US CMMS Medical Claims: bucket2009 classification
bdanalytics
Date: (Tue) Apr 21, 2015
Introduction:
Data: Source: Training: https://courses.edx.org/c4x/MITx/15.071x_2/asset/ClaimsData.csv.zip
New:
Time period:
Synopsis:
Based on analysis utilizing <> techniques,
[](.png)
Potential next steps include:
- Probability handling for multinomials vs. desired binomial outcome
- ROCR currently supports only evaluation of binary classification tasks (version 1.0.7)
extensions toward multiclass classification are scheduled for the next release
- Skip trControl.method=“cv” for dummy classifier ?
- Add custom model to caret for a dummy (baseline) classifier (binomial & multinomial) that generates proba/outcomes which mimics the freq distribution of glb_rsp_var values; Right now glb_dmy_glm_mdl always generates most frequent outcome in training data
glm_dmy_mdl should use the same method as glm_sel_mdl until custom dummy classifer is implemented
- Compare glb_sel_mdl vs. glb_fin_mdl:
- varImp
- Prediction differences (shd be minimal ?)
- Prediction accuracy scatter graph:
- Add tiles (raw vs. PCA)
- Use shiny for drop-down of “important” features
Use plot.ly for interactive plots ?
- Move glb_analytics_diag_plots to mydsutils.R: (+) Easier to debug (-) Too many glb vars used
- Add print(ggplot.petrinet(glb_analytics_pn) + coord_flip()) at the end of every major chunk
- Parameterize glb_analytics_pn
- Move glb_impute_missing_data to mydsutils.R: (-) Too many glb vars used; glb_<>_df reassigned
- Replicate myfit_mdl_classification features in myfit_mdl_regression
- Do non-glm methods handle interaction terms ?
- f-score computation for classifiers should be summation across outcomes (not just the desired one ?)
- Add accuracy computation to glb_dmy_mdl in predict.data.new chunk
- Why does splitting fit.data.training.all chunk into separate chunks add an overhead of ~30 secs ? It’s not rbind b/c other chunks have lower elapsed time. Is it the number of plots ?
- Incorporate code chunks in print_sessionInfo
- Test against
- projects in github.com/bdanalytics
- lectures in jhu-datascience track
Analysis:
rm(list=ls())
set.seed(12345)
options(stringsAsFactors=FALSE)
source("~/Dropbox/datascience/R/mydsutils.R")
source("~/Dropbox/datascience/R/myplot.R")
source("~/Dropbox/datascience/R/mypetrinet.R")
# Gather all package requirements here
#suppressPackageStartupMessages(require())
#packageVersion("caret")
#require(sos); findFn("pinv", maxPages=2, sortby="MaxScore")
# Analysis control global variables
glb_trnng_url <- "https://courses.edx.org/c4x/MITx/15.071x_2/asset/ClaimsData.csv.zip"
glb_newdt_url <- "<newdt_url>"
glb_is_separate_newent_dataset <- FALSE # or TRUE
glb_split_entity_newent_datasets <- TRUE # or FALSE
glb_split_newdata_method <- "sample" # "condition" or "sample"
glb_split_newdata_condition <- "<col_name> <condition_operator> <value>" # or NULL
glb_split_newdata_size_ratio <- 0.4 # > 0 & < 1
glb_split_sample.seed <- 88 # or any integer
glb_max_obs <- 500000 # or NULL
glb_is_regression <- FALSE; glb_is_classification <- TRUE
glb_rsp_var_raw <- "bucket2009"
# for classification, the response variable has to be a factor
# especially for random forests (method="rf")
glb_rsp_var <- "bucket2009.fctr" # or glb_rsp_var_raw
# if the response factor is based on numbers e.g (0/1 vs. "A"/"B"),
# caret predict(..., type="prob") crashes
glb_map_rsp_raw_to_var <- function(raw) {
as.factor(paste0("B", raw))
} # or NULL
#glb_map_rsp_raw_to_var(c(1, 2, 3, 4, 5))
glb_map_rsp_var_to_raw <- function(var) {
as.numeric(var)
} # or NULL
#glb_map_rsp_var_to_raw(glb_map_rsp_raw_to_var(c(1, 2, 3, 4, 5)))
if ((glb_rsp_var != glb_rsp_var_raw) & is.null(glb_map_rsp_raw_to_var))
stop("glb_map_rsp_raw_to_var function expected")
glb_rsp_var_out <- paste0(glb_rsp_var, ".predict.") # model_id is appended later
glb_id_vars <- NULL # or c("<id_var>")
# List transformed vars
glb_exclude_vars_as_features <- c("bucket2008.fctr") # or c(NULL)
# List feats that shd be excluded due to known causation by prediction variable
if (glb_rsp_var_raw != glb_rsp_var)
glb_exclude_vars_as_features <- union(glb_exclude_vars_as_features,
glb_rsp_var_raw)
glb_exclude_vars_as_features <- union(glb_exclude_vars_as_features,
c("reimbursement2009")) # or NULL
# List output vars (useful during testing in console)
# glb_exclude_vars_as_features <- union(glb_exclude_vars_as_features,
# grep(glb_rsp_var_out, names(glb_entity_df), value=TRUE))
glb_impute_na_data <- FALSE # or TRUE
glb_mice_complete.seed <- 144 # or any integer
# Classification
# rpart: .rnorm messes with the models badly
# caret creates dummy vars for factor feats which messes up the tuning
# - better to feed as.numeric(<feat>.fctr) to caret
#glb_models_method_vctr <- c("glm", "rpart", "rf") # Binomials
glb_models_method_vctr <- c("rpart", "rf") # Multinomials
#glb_models_method_vctr <- c("rpart") # Multinomials - this exercise
glb_models_lst <- list(); glb_models_df <- data.frame()
# Baseline prediction model feature(s)
glb_Baseline_mdl_var <- c("bucket2008.fctr") # or NULL
glb_model_metric_terms <- matrix(c(
0,1,2,3,4,
2,0,1,2,3,
4,2,0,1,2,
6,4,2,0,1,
8,6,4,2,0
), byrow=TRUE, nrow=5) # or NULL
glb_model_metric <- "loss.error" # or NULL
glb_model_metric_maximize <- FALSE # or NULL (TRUE is not the default for both classification & regression)
glb_model_metric_smmry <- function(data, lev=NULL, model=NULL) {
confusion_mtrx <- t(as.matrix(confusionMatrix(data$pred, data$obs)))
#print(confusion_mtrx)
#print(confusion_mtrx * glb_model_metric_terms)
metric <- sum(confusion_mtrx * glb_model_metric_terms) / nrow(data)
names(metric) <- glb_model_metric
return(metric)
}
glb_tune_models_df <-
rbind(
data.frame(parameter="cp", min=0.00005, max=0.00005, by=0.000005),
#min=0.00004; max=0.00006; by=0.000005
#data.frame(parameter="mtry", min=2, max=4, by=1),
data.frame(parameter="dummy", min=2, max=4, by=1)
)
# or NULL
glb_n_cv_folds <- 5 # or NULL
glb_clf_proba_threshold <- NULL # 0.5
# Model selection criteria
# For binomial classification add AIC
glb_model_sel_frmla <- formula(paste0("~ ",
ifelse(!is.null(glb_model_metric),
paste0(ifelse(!glb_model_metric_maximize, "+min.", "-max."),
paste0(glb_model_metric, ".OOB")),
""), " -max.Accuracy.OOB -max.Kappa.OOB"))
glb_sel_mdl_id <- "All.X.lser.ys.cp.4015.rpart" # or NULL
glb_fin_mdl_id <- glb_sel_mdl_id # or "Final"
# Depict process
glb_analytics_pn <- petrinet(name="glb_analytics_pn",
trans_df=data.frame(id=1:6,
name=c("data.training.all","data.new",
"model.selected","model.final",
"data.training.all.prediction","data.new.prediction"),
x=c( -5,-5,-15,-25,-25,-35),
y=c( -5, 5, 0, 0, -5, 5)
),
places_df=data.frame(id=1:4,
name=c("bgn","fit.data.training.all","predict.data.new","end"),
x=c( -0, -20, -30, -40),
y=c( 0, 0, 0, 0),
M0=c( 3, 0, 0, 0)
),
arcs_df=data.frame(
begin=c("bgn","bgn","bgn",
"data.training.all","model.selected","fit.data.training.all",
"fit.data.training.all","model.final",
"data.new","predict.data.new",
"data.training.all.prediction","data.new.prediction"),
end =c("data.training.all","data.new","model.selected",
"fit.data.training.all","fit.data.training.all","model.final",
"data.training.all.prediction","predict.data.new",
"predict.data.new","data.new.prediction",
"end","end")
))
#print(ggplot.petrinet(glb_analytics_pn))
print(ggplot.petrinet(glb_analytics_pn) + coord_flip())## Loading required package: grid
glb_analytics_avl_objs <- NULL
glb_script_tm <- proc.time()
glb_script_df <- data.frame(chunk_label="import_data",
chunk_step_major=1, chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"])
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed import_data 1 0 0.003Step 1: import data
glb_entity_df <- myimport_data(
url=glb_trnng_url,
comment="glb_entity_df", force_header=TRUE,
print_diagn=(glb_is_separate_newent_dataset |
!glb_split_entity_newent_datasets))## [1] "Reading file ./data/ClaimsData.csv..."
## [1] "dimensions of data in ./data/ClaimsData.csv: 458,005 rows x 16 cols"if (glb_is_separate_newent_dataset) {
glb_newent_df <- myimport_data(
url=glb_newdt_url,
comment="glb_newent_df", force_header=TRUE, print_diagn=TRUE)
} else {
if (!glb_split_entity_newent_datasets) {
stop("Not implemented yet")
glb_newent_df <- glb_entity_df[sample(1:nrow(glb_entity_df),
max(2, nrow(glb_entity_df) / 1000)),]
} else if (glb_split_newdata_method == "condition") {
glb_newent_df <- do.call("subset",
list(glb_entity_df, parse(text=glb_split_newdata_condition)))
glb_entity_df <- do.call("subset",
list(glb_entity_df, parse(text=paste0("!(",
glb_split_newdata_condition,
")"))))
} else if (glb_split_newdata_method == "sample") {
require(caTools)
set.seed(glb_split_sample.seed)
split <- sample.split(glb_entity_df[, glb_rsp_var_raw],
SplitRatio=(1-glb_split_newdata_size_ratio))
glb_newent_df <- glb_entity_df[!split, ]
glb_entity_df <- glb_entity_df[split ,]
} else stop("glb_split_newdata_method should be %in% c('condition', 'sample')")
comment(glb_newent_df) <- "glb_newent_df"
myprint_df(glb_newent_df)
str(glb_newent_df)
if (glb_split_entity_newent_datasets) {
myprint_df(glb_entity_df)
str(glb_entity_df)
}
} ## Loading required package: caTools## age alzheimers arthritis cancer copd depression diabetes heart.failure
## 3 67 0 0 0 0 0 0 0
## 5 67 0 0 0 0 0 0 0
## 6 68 0 0 0 0 0 0 0
## 8 70 0 0 0 0 0 0 0
## 9 67 0 0 0 0 0 0 0
## 10 67 0 0 0 0 0 0 0
## ihd kidney osteoporosis stroke reimbursement2008 bucket2008
## 3 0 0 0 0 0 1
## 5 0 0 0 0 0 1
## 6 0 0 0 0 0 1
## 8 0 0 0 0 0 1
## 9 0 0 0 0 0 1
## 10 0 0 0 0 0 1
## reimbursement2009 bucket2009
## 3 0 1
## 5 0 1
## 6 0 1
## 8 0 1
## 9 0 1
## 10 0 1
## age alzheimers arthritis cancer copd depression diabetes
## 43967 57 0 0 0 0 0 0
## 70246 70 0 0 0 0 0 0
## 165755 78 0 0 0 0 0 0
## 208131 73 0 1 1 0 0 0
## 319113 87 0 0 0 0 1 1
## 446073 72 1 0 1 0 1 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 43967 0 0 0 0 0 0
## 70246 0 0 0 0 0 0
## 165755 0 0 0 0 0 140
## 208131 0 0 0 1 0 5680
## 319113 0 1 0 0 0 2800
## 446073 1 1 0 1 1 16030
## bucket2008 reimbursement2009 bucket2009
## 43967 1 0 1
## 70246 1 0 1
## 165755 1 720 1
## 208131 2 1250 1
## 319113 1 3330 2
## 446073 3 28000 4
## age alzheimers arthritis cancer copd depression diabetes
## 457996 60 0 1 0 1 1 1
## 457998 87 0 0 0 1 1 1
## 458001 61 1 0 0 1 1 1
## 458002 90 1 0 0 1 1 1
## 458003 76 0 1 0 1 1 1
## 458005 80 1 0 0 1 1 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 457996 1 1 0 1 1 11720
## 457998 1 1 1 0 0 27750
## 458001 1 1 1 1 1 15960
## 458002 1 1 1 0 0 26870
## 458003 1 1 1 1 1 89140
## 458005 1 1 1 0 1 38320
## bucket2008 reimbursement2009 bucket2009
## 457996 3 142960 5
## 457998 4 148600 5
## 458001 3 154000 5
## 458002 4 155010 5
## 458003 5 155810 5
## 458005 4 189930 5
## 'data.frame': 183202 obs. of 16 variables:
## $ age : int 67 67 68 70 67 67 56 48 99 68 ...
## $ alzheimers : int 0 0 0 0 0 0 0 0 0 0 ...
## $ arthritis : int 0 0 0 0 0 0 0 0 0 0 ...
## $ cancer : int 0 0 0 0 0 0 0 0 0 0 ...
## $ copd : int 0 0 0 0 0 0 0 0 0 0 ...
## $ depression : int 0 0 0 0 0 0 0 0 0 0 ...
## $ diabetes : int 0 0 0 0 0 0 0 0 0 0 ...
## $ heart.failure : int 0 0 0 0 0 0 0 0 0 0 ...
## $ ihd : int 0 0 0 0 0 0 0 0 0 0 ...
## $ kidney : int 0 0 0 0 0 0 0 0 0 0 ...
## $ osteoporosis : int 0 0 0 0 0 0 0 0 0 0 ...
## $ stroke : int 0 0 0 0 0 0 0 0 0 0 ...
## $ reimbursement2008: int 0 0 0 0 0 0 0 0 0 0 ...
## $ bucket2008 : int 1 1 1 1 1 1 1 1 1 1 ...
## $ reimbursement2009: int 0 0 0 0 0 0 0 0 0 0 ...
## $ bucket2009 : int 1 1 1 1 1 1 1 1 1 1 ...
## - attr(*, "comment")= chr "glb_newent_df"
## age alzheimers arthritis cancer copd depression diabetes heart.failure
## 1 85 0 0 0 0 0 0 0
## 2 59 0 0 0 0 0 0 0
## 4 52 0 0 0 0 0 0 0
## 7 75 0 0 0 0 0 0 0
## 11 89 0 0 0 0 0 0 0
## 13 74 0 0 0 0 0 0 0
## ihd kidney osteoporosis stroke reimbursement2008 bucket2008
## 1 0 0 0 0 0 1
## 2 0 0 0 0 0 1
## 4 0 0 0 0 0 1
## 7 0 0 0 0 0 1
## 11 0 0 0 0 0 1
## 13 0 0 0 0 0 1
## reimbursement2009 bucket2009
## 1 0 1
## 2 0 1
## 4 0 1
## 7 0 1
## 11 0 1
## 13 0 1
## age alzheimers arthritis cancer copd depression diabetes
## 138659 69 0 0 0 0 0 0
## 168428 74 1 0 0 0 0 1
## 189703 81 0 0 0 0 0 1
## 225640 78 1 0 0 0 1 0
## 382169 77 1 0 0 1 1 1
## 397881 46 1 0 0 0 0 0
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 138659 0 0 0 0 0 0
## 168428 0 0 0 1 0 720
## 189703 0 1 0 0 0 690
## 225640 0 0 0 1 0 1540
## 382169 1 1 1 1 1 16400
## 397881 1 1 1 0 0 3700
## bucket2008 reimbursement2009 bucket2009
## 138659 1 380 1
## 168428 1 750 1
## 189703 1 1020 1
## 225640 1 1490 1
## 382169 3 6620 2
## 397881 2 8470 3
## age alzheimers arthritis cancer copd depression diabetes
## 457991 76 0 0 0 1 1 1
## 457992 84 0 0 0 1 0 1
## 457997 73 0 0 0 1 1 1
## 457999 83 1 1 0 1 0 1
## 458000 56 0 1 0 1 1 1
## 458004 82 1 0 0 1 0 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 457991 1 1 1 1 0 53550
## 457992 1 1 1 0 0 8620
## 457997 1 1 1 1 0 53230
## 457999 1 1 1 1 1 62620
## 458000 1 1 1 1 0 62980
## 458004 1 1 1 1 1 20660
## bucket2008 reimbursement2009 bucket2009
## 457991 4 131960 5
## 457992 3 133500 5
## 457997 4 147760 5
## 457999 5 148860 5
## 458000 5 151880 5
## 458004 4 158800 5
## 'data.frame': 274803 obs. of 16 variables:
## $ age : int 85 59 52 75 89 74 81 86 78 67 ...
## $ alzheimers : int 0 0 0 0 0 0 0 0 0 0 ...
## $ arthritis : int 0 0 0 0 0 0 0 0 0 0 ...
## $ cancer : int 0 0 0 0 0 0 0 0 0 0 ...
## $ copd : int 0 0 0 0 0 0 0 0 0 0 ...
## $ depression : int 0 0 0 0 0 0 0 0 0 0 ...
## $ diabetes : int 0 0 0 0 0 0 0 0 0 0 ...
## $ heart.failure : int 0 0 0 0 0 0 0 0 0 0 ...
## $ ihd : int 0 0 0 0 0 0 0 0 0 0 ...
## $ kidney : int 0 0 0 0 0 0 0 0 0 0 ...
## $ osteoporosis : int 0 0 0 0 0 0 0 0 0 0 ...
## $ stroke : int 0 0 0 0 0 0 0 0 0 0 ...
## $ reimbursement2008: int 0 0 0 0 0 0 0 0 0 0 ...
## $ bucket2008 : int 1 1 1 1 1 1 1 1 1 1 ...
## $ reimbursement2009: int 0 0 0 0 0 0 0 0 0 0 ...
## $ bucket2009 : int 1 1 1 1 1 1 1 1 1 1 ...
## - attr(*, "comment")= chr "glb_entity_df"if (!is.null(glb_max_obs)) {
if (nrow(glb_entity_df) > glb_max_obs) {
warning("glb_entity_df restricted to glb_max_obs: ", format(glb_max_obs, big.mark=","))
org_entity_df <- glb_entity_df
glb_entity_df <- org_entity_df[split <-
sample.split(org_entity_df[, glb_rsp_var_raw], SplitRatio=glb_max_obs), ]
org_entity_df <- NULL
}
if (nrow(glb_newent_df) > glb_max_obs) {
warning("glb_newent_df restricted to glb_max_obs: ", format(glb_max_obs, big.mark=","))
org_newent_df <- glb_newent_df
glb_newent_df <- org_newent_df[split <-
sample.split(org_newent_df[, glb_rsp_var_raw], SplitRatio=glb_max_obs), ]
org_newent_df <- NULL
}
}
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="cleanse_data",
chunk_step_major=max(glb_script_df$chunk_step_major)+1,
chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed import_data 1 0 0.003
## elapsed1 cleanse_data 2 0 7.036Step 2: cleanse data
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="inspectORexplore.data",
chunk_step_major=max(glb_script_df$chunk_step_major),
chunk_step_minor=1,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed1 cleanse_data 2 0 7.036
## elapsed2 inspectORexplore.data 2 1 7.066Step 2.1: inspect/explore data
#print(str(glb_entity_df))
#View(glb_entity_df)
# List info gathered for various columns
# <col_name>: <description>; <notes>
# Create new features that help diagnostics
# Create factors of string variables
str_vars <- sapply(1:length(names(glb_entity_df)),
function(col) ifelse(class(glb_entity_df[, names(glb_entity_df)[col]]) == "character",
names(glb_entity_df)[col], ""))
if (length(str_vars <- setdiff(str_vars[str_vars != ""],
glb_exclude_vars_as_features)) > 0) {
warning("Creating factors of string variables:", paste0(str_vars, collapse=", "))
glb_exclude_vars_as_features <- union(glb_exclude_vars_as_features, str_vars)
for (var in str_vars) {
glb_entity_df[, paste0(var, ".fctr")] <- factor(glb_entity_df[, var],
as.factor(union(glb_entity_df[, var], glb_newent_df[, var])))
glb_newent_df[, paste0(var, ".fctr")] <- factor(glb_newent_df[, var],
as.factor(union(glb_entity_df[, var], glb_newent_df[, var])))
}
}
# Convert factors to dummy variables
# Build splines require(splines); bsBasis <- bs(training$age, df=3)
add_new_diag_feats <- function(obs_df, obs_twin_df) {
require(plyr)
obs_df <- mutate(obs_df,
# <col_name>.NA=is.na(<col_name>),
# <col_name>.fctr=factor(<col_name>,
# as.factor(union(obs_df$<col_name>, obs_twin_df$<col_name>))),
# <col_name>.fctr=relevel(factor(<col_name>,
# as.factor(union(obs_df$<col_name>, obs_twin_df$<col_name>))),
# "<ref_val>"),
# <col2_name>.fctr=relevel(factor(ifelse(<col1_name> == <val>, "<oth_val>", "<ref_val>")),
# as.factor(c("R", "<ref_val>")),
# ref="<ref_val>"),
# This doesn't work - use sapply instead
# <col_name>.fctr_num=grep(<col_name>, levels(<col_name>.fctr)),
#
# Date.my=as.Date(strptime(Date, "%m/%d/%y %H:%M")),
# Year=year(Date.my),
# Month=months(Date.my),
# Weekday=weekdays(Date.my)
# <col_name>.log=log(<col.name>),
# <col_name>=<table>[as.character(<col2_name>)],
# <col_name>=as.numeric(<col2_name>),
.rnorm=rnorm(n=nrow(obs_df))
)
# If levels of a factor are different across obs_df & glb_newent_df; predict.glm fails
# Transformations not handled by mutate
# obs_df$<col_name>.fctr.num <- sapply(1:nrow(obs_df),
# function(row_ix) grep(obs_df[row_ix, "<col_name>"],
# levels(obs_df[row_ix, "<col_name>.fctr"])))
print(summary(obs_df))
print(sapply(names(obs_df), function(col) sum(is.na(obs_df[, col]))))
return(obs_df)
}
glb_entity_df <- add_new_diag_feats(glb_entity_df, glb_newent_df)## Loading required package: plyr## age alzheimers arthritis cancer
## Min. : 26.00 Min. :0.0000 Min. :0.0000 Min. :0.00000
## 1st Qu.: 67.00 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000
## Median : 73.00 Median :0.0000 Median :0.0000 Median :0.00000
## Mean : 72.64 Mean :0.1928 Mean :0.1546 Mean :0.06402
## 3rd Qu.: 81.00 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.00000
## Max. :100.00 Max. :1.0000 Max. :1.0000 Max. :1.00000
## copd depression diabetes heart.failure
## Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.1369 Mean :0.2129 Mean :0.3809 Mean :0.2852
## 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:1.0000
## Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000
## ihd kidney osteoporosis stroke
## Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.00000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000
## Median :0.0000 Median :0.0000 Median :0.0000 Median :0.00000
## Mean :0.4205 Mean :0.1616 Mean :0.1738 Mean :0.04497
## 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.00000
## Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.00000
## reimbursement2008 bucket2008 reimbursement2009 bucket2009
## Min. : 0 Min. :1.000 Min. : 0 Min. :1.000
## 1st Qu.: 0 1st Qu.:1.000 1st Qu.: 130 1st Qu.:1.000
## Median : 960 Median :1.000 Median : 1540 Median :1.000
## Mean : 4016 Mean :1.438 Mean : 4280 Mean :1.522
## 3rd Qu.: 3110 3rd Qu.:2.000 3rd Qu.: 4220 3rd Qu.:2.000
## Max. :194910 Max. :5.000 Max. :158800 Max. :5.000
## .rnorm
## Min. :-4.627678
## 1st Qu.:-0.668424
## Median : 0.002423
## Mean : 0.001867
## 3rd Qu.: 0.675532
## Max. : 4.385255
## age alzheimers arthritis cancer
## 0 0 0 0
## copd depression diabetes heart.failure
## 0 0 0 0
## ihd kidney osteoporosis stroke
## 0 0 0 0
## reimbursement2008 bucket2008 reimbursement2009 bucket2009
## 0 0 0 0
## .rnorm
## 0glb_newent_df <- add_new_diag_feats(glb_newent_df, glb_entity_df)## age alzheimers arthritis cancer
## Min. : 26.00 Min. :0.0000 Min. :0.0000 Min. :0.00000
## 1st Qu.: 67.00 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000
## Median : 73.00 Median :0.0000 Median :0.0000 Median :0.00000
## Mean : 72.61 Mean :0.1914 Mean :0.1539 Mean :0.06424
## 3rd Qu.: 81.00 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.00000
## Max. :100.00 Max. :1.0000 Max. :1.0000 Max. :1.00000
## copd depression diabetes heart.failure
## Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.1348 Mean :0.2133 Mean :0.3798 Mean :0.2841
## 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:1.0000
## Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000
## ihd kidney osteoporosis stroke
## Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.00000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000
## Median :0.0000 Median :0.0000 Median :0.0000 Median :0.00000
## Mean :0.4195 Mean :0.1605 Mean :0.1743 Mean :0.04453
## 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.00000
## Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.00000
## reimbursement2008 bucket2008 reimbursement2009 bucket2009
## Min. : 0 Min. :1.000 Min. : 0 Min. :1.000
## 1st Qu.: 0 1st Qu.:1.000 1st Qu.: 120 1st Qu.:1.000
## Median : 950 Median :1.000 Median : 1530 Median :1.000
## Mean : 3989 Mean :1.436 Mean : 4273 Mean :1.522
## 3rd Qu.: 3110 3rd Qu.:2.000 3rd Qu.: 4210 3rd Qu.:2.000
## Max. :221640 Max. :5.000 Max. :189930 Max. :5.000
## .rnorm
## Min. :-4.224787
## 1st Qu.:-0.673311
## Median : 0.001064
## Mean : 0.002116
## 3rd Qu.: 0.679841
## Max. : 4.235888
## age alzheimers arthritis cancer
## 0 0 0 0
## copd depression diabetes heart.failure
## 0 0 0 0
## ihd kidney osteoporosis stroke
## 0 0 0 0
## reimbursement2008 bucket2008 reimbursement2009 bucket2009
## 0 0 0 0
## .rnorm
## 0# Histogram of predictor in glb_entity_df & glb_newent_df
print(myplot_histogram(glb_entity_df, glb_rsp_var_raw))## stat_bin: binwidth defaulted to range/30. Use 'binwidth = x' to adjust this.
if (glb_is_classification)
print(table(glb_entity_df[, glb_rsp_var_raw]) / nrow(glb_entity_df))##
## 1 2 3 4 5
## 0.671266325 0.190168957 0.089467728 0.043325582 0.005771407print(myplot_histogram(glb_newent_df, glb_rsp_var_raw))## stat_bin: binwidth defaulted to range/30. Use 'binwidth = x' to adjust this.
# Check for duplicates in glb_id_vars
if (length(glb_id_vars) > 0) {
id_vars_dups_df <- subset(id_vars_df <- mycreate_tbl_df(
rbind(glb_entity_df[, glb_id_vars, FALSE], glb_newent_df[, glb_id_vars, FALSE]),
glb_id_vars), .freq > 1)
if (nrow(id_vars_dups_df) > 0) {
warning("Duplicates found in glb_id_vars data:", nrow(id_vars_dups_df))
myprint_df(id_vars_dups_df)
} else {
# glb_id_vars are unique across obs in both glb_<>_df
glb_exclude_vars_as_features <- union(glb_exclude_vars_as_features, glb_id_vars)
}
}
#pairs(subset(glb_entity_df, select=-c(col_symbol)))
# Check for glb_newent_df & glb_entity_df features range mismatches
# Other diagnostics:
# print(subset(glb_entity_df, <col1_name> == max(glb_entity_df$<col1_name>, na.rm=TRUE) &
# <col2_name> <= mean(glb_entity_df$<col1_name>, na.rm=TRUE)))
# print(glb_entity_df[which.max(glb_entity_df$<col_name>),])
# print(<col_name>_freq_glb_entity_df <- mycreate_tbl_df(glb_entity_df, "<col_name>"))
# print(which.min(table(glb_entity_df$<col_name>)))
# print(which.max(table(glb_entity_df$<col_name>)))
# print(which.max(table(glb_entity_df$<col1_name>, glb_entity_df$<col2_name>)[, 2]))
# print(table(glb_entity_df$<col1_name>, glb_entity_df$<col2_name>))
# print(table(is.na(glb_entity_df$<col1_name>), glb_entity_df$<col2_name>))
# print(table(sign(glb_entity_df$<col1_name>), glb_entity_df$<col2_name>))
# print(mycreate_xtab(glb_entity_df, <col1_name>))
# print(mycreate_xtab(glb_entity_df, c(<col1_name>, <col2_name>)))
# print(<col1_name>_<col2_name>_xtab_glb_entity_df <-
# mycreate_xtab(glb_entity_df, c("<col1_name>", "<col2_name>")))
# <col1_name>_<col2_name>_xtab_glb_entity_df[is.na(<col1_name>_<col2_name>_xtab_glb_entity_df)] <- 0
# print(<col1_name>_<col2_name>_xtab_glb_entity_df <-
# mutate(<col1_name>_<col2_name>_xtab_glb_entity_df,
# <col3_name>=(<col1_name> * 1.0) / (<col1_name> + <col2_name>)))
# print(<col2_name>_min_entity_arr <-
# sort(tapply(glb_entity_df$<col1_name>, glb_entity_df$<col2_name>, min, na.rm=TRUE)))
# print(<col1_name>_na_by_<col2_name>_arr <-
# sort(tapply(glb_entity_df$<col1_name>.NA, glb_entity_df$<col2_name>, mean, na.rm=TRUE)))
# Other plots:
# print(myplot_box(df=glb_entity_df, ycol_names="<col1_name>"))
# print(myplot_box(df=glb_entity_df, ycol_names="<col1_name>", xcol_name="<col2_name>"))
# print(myplot_line(subset(glb_entity_df, Symbol %in% c("KO", "PG")),
# "Date.my", "StockPrice", facet_row_colnames="Symbol") +
# geom_vline(xintercept=as.numeric(as.Date("2003-03-01"))) +
# geom_vline(xintercept=as.numeric(as.Date("1983-01-01")))
# )
# print(myplot_scatter(glb_entity_df, "<col1_name>", "<col2_name>", smooth=TRUE))
# print(myplot_scatter(glb_entity_df, "<col1_name>", "<col2_name>", colorcol_name="<Pred.fctr>"))
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="manage_missing_data",
chunk_step_major=max(glb_script_df$chunk_step_major),
chunk_step_minor=glb_script_df[nrow(glb_script_df), "chunk_step_minor"]+1,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed2 inspectORexplore.data 2 1 7.066
## elapsed3 manage_missing_data 2 2 16.418Step 2.2: manage missing data
# print(sapply(names(glb_entity_df), function(col) sum(is.na(glb_entity_df[, col]))))
# print(sapply(names(glb_newent_df), function(col) sum(is.na(glb_newent_df[, col]))))
# glb_entity_df <- na.omit(glb_entity_df)
# glb_newent_df <- na.omit(glb_newent_df)
# df[is.na(df)] <- 0
# Not refactored into mydsutils.R since glb_*_df might be reassigned
glb_impute_missing_data <- function(entity_df, newent_df) {
if (!glb_is_separate_newent_dataset) {
# Combine entity & newent
union_df <- rbind(mutate(entity_df, .src = "entity"),
mutate(newent_df, .src = "newent"))
union_imputed_df <- union_df[, setdiff(setdiff(names(entity_df),
glb_rsp_var),
glb_exclude_vars_as_features)]
print(summary(union_imputed_df))
require(mice)
set.seed(glb_mice_complete.seed)
union_imputed_df <- complete(mice(union_imputed_df))
print(summary(union_imputed_df))
union_df[, names(union_imputed_df)] <- union_imputed_df[, names(union_imputed_df)]
print(summary(union_df))
# union_df$.rownames <- rownames(union_df)
# union_df <- orderBy(~.rownames, union_df)
#
# imp_entity_df <- myimport_data(
# url="<imputed_trnng_url>",
# comment="imp_entity_df", force_header=TRUE, print_diagn=TRUE)
# print(all.equal(subset(union_df, select=-c(.src, .rownames, .rnorm)),
# imp_entity_df))
# Partition again
glb_entity_df <<- subset(union_df, .src == "entity", select=-c(.src, .rownames))
comment(glb_entity_df) <- "entity_df"
glb_newent_df <<- subset(union_df, .src == "newent", select=-c(.src, .rownames))
comment(glb_newent_df) <- "newent_df"
# Generate summaries
print(summary(entity_df))
print(sapply(names(entity_df), function(col) sum(is.na(entity_df[, col]))))
print(summary(newent_df))
print(sapply(names(newent_df), function(col) sum(is.na(newent_df[, col]))))
} else stop("Not implemented yet")
}
if (glb_impute_na_data) {
if ((sum(sapply(names(glb_entity_df),
function(col) sum(is.na(glb_entity_df[, col])))) > 0) |
(sum(sapply(names(glb_newent_df),
function(col) sum(is.na(glb_newent_df[, col])))) > 0))
glb_impute_missing_data(glb_entity_df, glb_newent_df)
}
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="encode_retype_data",
chunk_step_major=max(glb_script_df$chunk_step_major),
chunk_step_minor=glb_script_df[nrow(glb_script_df), "chunk_step_minor"]+1,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed3 manage_missing_data 2 2 16.418
## elapsed4 encode_retype_data 2 3 16.992Step 2.3: encode/retype data
# map_<col_name>_df <- myimport_data(
# url="<map_url>",
# comment="map_<col_name>_df", print_diagn=TRUE)
# map_<col_name>_df <- read.csv(paste0(getwd(), "/data/<file_name>.csv"), strip.white=TRUE)
# glb_entity_df <- mymap_codes(glb_entity_df, "<from_col_name>", "<to_col_name>",
# map_<to_col_name>_df, map_join_col_name="<map_join_col_name>",
# map_tgt_col_name="<to_col_name>")
# glb_newent_df <- mymap_codes(glb_newent_df, "<from_col_name>", "<to_col_name>",
# map_<to_col_name>_df, map_join_col_name="<map_join_col_name>",
# map_tgt_col_name="<to_col_name>")
# glb_entity_df$<col_name>.fctr <- factor(glb_entity_df$<col_name>,
# as.factor(union(glb_entity_df$<col_name>, glb_newent_df$<col_name>)))
# glb_newent_df$<col_name>.fctr <- factor(glb_newent_df$<col_name>,
# as.factor(union(glb_entity_df$<col_name>, glb_newent_df$<col_name>)))
if (!is.null(glb_map_rsp_raw_to_var)) {
glb_entity_df[, glb_rsp_var] <-
glb_map_rsp_raw_to_var(glb_entity_df[, glb_rsp_var_raw])
mycheck_map_results(mapd_df=glb_entity_df,
from_col_name=glb_rsp_var_raw, to_col_name=glb_rsp_var)
glb_newent_df[, glb_rsp_var] <-
glb_map_rsp_raw_to_var(glb_newent_df[, glb_rsp_var_raw])
mycheck_map_results(mapd_df=glb_newent_df,
from_col_name=glb_rsp_var_raw, to_col_name=glb_rsp_var)
glb_entity_df[, "bucket2008.fctr"] <-
glb_map_rsp_raw_to_var(glb_entity_df[, "bucket2008"])
mycheck_map_results(mapd_df=glb_entity_df,
from_col_name="bucket2008", to_col_name="bucket2008.fctr")
glb_newent_df[, "bucket2008.fctr"] <-
glb_map_rsp_raw_to_var(glb_newent_df[, "bucket2008"])
mycheck_map_results(mapd_df=glb_newent_df,
from_col_name="bucket2008", to_col_name="bucket2008.fctr")
}## Loading required package: sqldf
## Loading required package: gsubfn
## Loading required package: proto
## Loading required package: RSQLite
## Loading required package: DBI
## Loading required package: tcltk## bucket2009 bucket2009.fctr .n
## 1 1 B1 184466
## 2 2 B2 52259
## 3 3 B3 24586
## 4 4 B4 11906
## 5 5 B5 1586
## bucket2009 bucket2009.fctr .n
## 1 1 B1 122978
## 2 2 B2 34840
## 3 3 B3 16390
## 4 4 B4 7937
## 5 5 B5 1057
## bucket2008 bucket2008.fctr .n
## 1 1 B1 204077
## 2 2 B2 37902
## 3 3 B3 18442
## 4 4 B4 12062
## 5 5 B5 2320
## bucket2008 bucket2008.fctr .n
## 1 1 B1 136125
## 2 2 B2 25271
## 3 3 B3 12405
## 4 4 B4 7852
## 5 5 B5 1549
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="extract_features",
chunk_step_major=max(glb_script_df$chunk_step_major)+1,
chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed4 encode_retype_data 2 3 16.992
## elapsed5 extract_features 3 0 26.774Step 3: extract features
# Create new features that help prediction
# <col_name>.lag.2 <- lag(zoo(glb_entity_df$<col_name>), -2, na.pad=TRUE)
# glb_entity_df[, "<col_name>.lag.2"] <- coredata(<col_name>.lag.2)
# <col_name>.lag.2 <- lag(zoo(glb_newent_df$<col_name>), -2, na.pad=TRUE)
# glb_newent_df[, "<col_name>.lag.2"] <- coredata(<col_name>.lag.2)
#
# glb_newent_df[1, "<col_name>.lag.2"] <- glb_entity_df[nrow(glb_entity_df) - 1,
# "<col_name>"]
# glb_newent_df[2, "<col_name>.lag.2"] <- glb_entity_df[nrow(glb_entity_df),
# "<col_name>"]
# glb_entity_df <- mutate(glb_entity_df,
# <new_col_name>=
# )
# glb_newent_df <- mutate(glb_newent_df,
# <new_col_name>=
# )
# print(summary(glb_entity_df))
# print(summary(glb_newent_df))
# print(sapply(names(glb_entity_df), function(col) sum(is.na(glb_entity_df[, col]))))
# print(sapply(names(glb_newent_df), function(col) sum(is.na(glb_newent_df[, col]))))
# print(myplot_scatter(glb_entity_df, "<col1_name>", "<col2_name>", smooth=TRUE))
replay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"data.training.all","data.new")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="select_features",
chunk_step_major=max(glb_script_df$chunk_step_major)+1,
chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed5 extract_features 3 0 26.774
## elapsed6 select_features 4 0 28.472Step 4: select features
print(glb_feats_df <- myselect_features(entity_df=glb_entity_df,
exclude_vars_as_features=glb_exclude_vars_as_features,
rsp_var=glb_rsp_var))## id cor.y exclude.as.feat
## bucket2009 bucket2009 1.0000000000 1
## reimbursement2009 reimbursement2009 0.8581631864 1
## bucket2008 bucket2008 0.4518935763 0
## bucket2008.fctr bucket2008.fctr 0.4518935763 1
## ihd ihd 0.3905905884 0
## diabetes diabetes 0.3904719536 0
## reimbursement2008 reimbursement2008 0.3756473063 0
## kidney kidney 0.3683780944 0
## heart.failure heart.failure 0.3647689526 0
## copd copd 0.3108325355 0
## depression depression 0.2835366153 0
## alzheimers alzheimers 0.2741643394 0
## arthritis arthritis 0.2717113526 0
## cancer cancer 0.2100892954 0
## osteoporosis osteoporosis 0.2076745377 0
## stroke stroke 0.1846626746 0
## age age 0.0495694151 0
## .rnorm .rnorm -0.0007970401 0
## cor.y.abs
## bucket2009 1.0000000000
## reimbursement2009 0.8581631864
## bucket2008 0.4518935763
## bucket2008.fctr 0.4518935763
## ihd 0.3905905884
## diabetes 0.3904719536
## reimbursement2008 0.3756473063
## kidney 0.3683780944
## heart.failure 0.3647689526
## copd 0.3108325355
## depression 0.2835366153
## alzheimers 0.2741643394
## arthritis 0.2717113526
## cancer 0.2100892954
## osteoporosis 0.2076745377
## stroke 0.1846626746
## age 0.0495694151
## .rnorm 0.0007970401glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="remove_correlated_features",
chunk_step_major=max(glb_script_df$chunk_step_major),
chunk_step_minor=glb_script_df[nrow(glb_script_df), "chunk_step_minor"]+1,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor
## elapsed6 select_features 4 0
## elapsed7 remove_correlated_features 4 1
## elapsed
## elapsed6 28.472
## elapsed7 29.537Step 4.1: remove correlated features
print(glb_feats_df <- orderBy(~-cor.y,
myfind_cor_features(feats_df=glb_feats_df, entity_df=glb_entity_df,
rsp_var=glb_rsp_var)))## Loading required package: reshape2## bucket2008 ihd diabetes reimbursement2008
## bucket2008 1.000000000 0.457745311 0.441695494 0.847360100
## ihd 0.457745311 1.000000000 0.514681463 0.362167974
## diabetes 0.441695494 0.514681463 1.000000000 0.349826333
## reimbursement2008 0.847360100 0.362167974 0.349826333 1.000000000
## kidney 0.556443060 0.379348673 0.410508010 0.482661680
## heart.failure 0.483790416 0.466853617 0.447713361 0.397862155
## copd 0.479911725 0.341479503 0.337436848 0.411894455
## depression 0.342770130 0.323337325 0.347401660 0.275910696
## alzheimers 0.375421748 0.326534346 0.341021686 0.313974193
## arthritis 0.330250474 0.311822976 0.323405716 0.250483735
## cancer 0.295879956 0.192569561 0.182463340 0.244638721
## osteoporosis 0.234597479 0.265948840 0.273431836 0.185149970
## stroke 0.319686629 0.193566819 0.191335431 0.296068271
## age 0.065957565 0.072806906 0.067414459 0.049189684
## .rnorm -0.005471367 -0.002117255 -0.001223323 -0.003838492
## kidney heart.failure copd depression
## bucket2008 0.556443060 0.48379042 0.479911725 0.342770130
## ihd 0.379348673 0.46685362 0.341479503 0.323337325
## diabetes 0.410508010 0.44771336 0.337436848 0.347401660
## reimbursement2008 0.482661680 0.39786215 0.411894455 0.275910696
## kidney 1.000000000 0.41853590 0.362145211 0.274868719
## heart.failure 0.418535903 1.00000000 0.376619956 0.303108016
## copd 0.362145211 0.37661996 1.000000000 0.262091958
## depression 0.274868719 0.30310802 0.262091958 1.000000000
## alzheimers 0.304331472 0.33163762 0.282316055 0.289283692
## arthritis 0.250350192 0.26577236 0.218711881 0.231412945
## cancer 0.187306219 0.17399979 0.173576229 0.120859944
## osteoporosis 0.192706719 0.22170825 0.184441404 0.193200212
## stroke 0.229542821 0.22049646 0.206759478 0.155228820
## age 0.056523917 0.06758002 0.044475985 0.013183881
## .rnorm -0.006698442 -0.00413497 -0.003209717 -0.004546531
## alzheimers arthritis cancer osteoporosis
## bucket2008 0.375421748 0.330250474 0.295879956 0.2345974786
## ihd 0.326534346 0.311822976 0.192569561 0.2659488405
## diabetes 0.341021686 0.323405716 0.182463340 0.2734318364
## reimbursement2008 0.313974193 0.250483735 0.244638721 0.1851499698
## kidney 0.304331472 0.250350192 0.187306219 0.1927067194
## heart.failure 0.331637621 0.265772357 0.173999790 0.2217082477
## copd 0.282316055 0.218711881 0.173576229 0.1844414035
## depression 0.289283692 0.231412945 0.120859944 0.1932002120
## alzheimers 1.000000000 0.205787238 0.136229360 0.1898304044
## arthritis 0.205787238 1.000000000 0.116224739 0.2188952357
## cancer 0.136229360 0.116224739 1.000000000 0.1145554395
## osteoporosis 0.189830404 0.218895236 0.114555439 1.0000000000
## stroke 0.217944953 0.117518638 0.100858117 0.1000348636
## age 0.044509085 0.042003908 0.037706744 0.0400590867
## .rnorm -0.002242699 -0.003328174 -0.001080007 -0.0007172342
## stroke age .rnorm
## bucket2008 0.319686629 0.0659575650 -0.0054713667
## ihd 0.193566819 0.0728069059 -0.0021172555
## diabetes 0.191335431 0.0674144586 -0.0012233233
## reimbursement2008 0.296068271 0.0491896842 -0.0038384924
## kidney 0.229542821 0.0565239171 -0.0066984421
## heart.failure 0.220496456 0.0675800164 -0.0041349700
## copd 0.206759478 0.0444759850 -0.0032097166
## depression 0.155228820 0.0131838813 -0.0045465313
## alzheimers 0.217944953 0.0445090852 -0.0022426985
## arthritis 0.117518638 0.0420039076 -0.0033281745
## cancer 0.100858117 0.0377067437 -0.0010800072
## osteoporosis 0.100034864 0.0400590867 -0.0007172342
## stroke 1.000000000 0.0365102875 -0.0034834136
## age 0.036510287 1.0000000000 0.0007606978
## .rnorm -0.003483414 0.0007606978 1.0000000000
## bucket2008 ihd diabetes reimbursement2008
## bucket2008 0.000000000 0.457745311 0.441695494 0.847360100
## ihd 0.457745311 0.000000000 0.514681463 0.362167974
## diabetes 0.441695494 0.514681463 0.000000000 0.349826333
## reimbursement2008 0.847360100 0.362167974 0.349826333 0.000000000
## kidney 0.556443060 0.379348673 0.410508010 0.482661680
## heart.failure 0.483790416 0.466853617 0.447713361 0.397862155
## copd 0.479911725 0.341479503 0.337436848 0.411894455
## depression 0.342770130 0.323337325 0.347401660 0.275910696
## alzheimers 0.375421748 0.326534346 0.341021686 0.313974193
## arthritis 0.330250474 0.311822976 0.323405716 0.250483735
## cancer 0.295879956 0.192569561 0.182463340 0.244638721
## osteoporosis 0.234597479 0.265948840 0.273431836 0.185149970
## stroke 0.319686629 0.193566819 0.191335431 0.296068271
## age 0.065957565 0.072806906 0.067414459 0.049189684
## .rnorm 0.005471367 0.002117255 0.001223323 0.003838492
## kidney heart.failure copd depression
## bucket2008 0.556443060 0.48379042 0.479911725 0.342770130
## ihd 0.379348673 0.46685362 0.341479503 0.323337325
## diabetes 0.410508010 0.44771336 0.337436848 0.347401660
## reimbursement2008 0.482661680 0.39786215 0.411894455 0.275910696
## kidney 0.000000000 0.41853590 0.362145211 0.274868719
## heart.failure 0.418535903 0.00000000 0.376619956 0.303108016
## copd 0.362145211 0.37661996 0.000000000 0.262091958
## depression 0.274868719 0.30310802 0.262091958 0.000000000
## alzheimers 0.304331472 0.33163762 0.282316055 0.289283692
## arthritis 0.250350192 0.26577236 0.218711881 0.231412945
## cancer 0.187306219 0.17399979 0.173576229 0.120859944
## osteoporosis 0.192706719 0.22170825 0.184441404 0.193200212
## stroke 0.229542821 0.22049646 0.206759478 0.155228820
## age 0.056523917 0.06758002 0.044475985 0.013183881
## .rnorm 0.006698442 0.00413497 0.003209717 0.004546531
## alzheimers arthritis cancer osteoporosis
## bucket2008 0.375421748 0.330250474 0.295879956 0.2345974786
## ihd 0.326534346 0.311822976 0.192569561 0.2659488405
## diabetes 0.341021686 0.323405716 0.182463340 0.2734318364
## reimbursement2008 0.313974193 0.250483735 0.244638721 0.1851499698
## kidney 0.304331472 0.250350192 0.187306219 0.1927067194
## heart.failure 0.331637621 0.265772357 0.173999790 0.2217082477
## copd 0.282316055 0.218711881 0.173576229 0.1844414035
## depression 0.289283692 0.231412945 0.120859944 0.1932002120
## alzheimers 0.000000000 0.205787238 0.136229360 0.1898304044
## arthritis 0.205787238 0.000000000 0.116224739 0.2188952357
## cancer 0.136229360 0.116224739 0.000000000 0.1145554395
## osteoporosis 0.189830404 0.218895236 0.114555439 0.0000000000
## stroke 0.217944953 0.117518638 0.100858117 0.1000348636
## age 0.044509085 0.042003908 0.037706744 0.0400590867
## .rnorm 0.002242699 0.003328174 0.001080007 0.0007172342
## stroke age .rnorm
## bucket2008 0.319686629 0.0659575650 0.0054713667
## ihd 0.193566819 0.0728069059 0.0021172555
## diabetes 0.191335431 0.0674144586 0.0012233233
## reimbursement2008 0.296068271 0.0491896842 0.0038384924
## kidney 0.229542821 0.0565239171 0.0066984421
## heart.failure 0.220496456 0.0675800164 0.0041349700
## copd 0.206759478 0.0444759850 0.0032097166
## depression 0.155228820 0.0131838813 0.0045465313
## alzheimers 0.217944953 0.0445090852 0.0022426985
## arthritis 0.117518638 0.0420039076 0.0033281745
## cancer 0.100858117 0.0377067437 0.0010800072
## osteoporosis 0.100034864 0.0400590867 0.0007172342
## stroke 0.000000000 0.0365102875 0.0034834136
## age 0.036510287 0.0000000000 0.0007606978
## .rnorm 0.003483414 0.0007606978 0.0000000000
## [1] "cor(bucket2008, reimbursement2008)=0.8474"
## [1] "cor(bucket2009.fctr, bucket2008)=0.4519"
## [1] "cor(bucket2009.fctr, reimbursement2008)=0.3756"## geom_smooth: Only one unique x value each group.Maybe you want aes(group = 1)?
## geom_smooth: Only one unique x value each group.Maybe you want aes(group = 1)?
## geom_smooth: Only one unique x value each group.Maybe you want aes(group = 1)?
## geom_smooth: Only one unique x value each group.Maybe you want aes(group = 1)?## Warning in myfind_cor_features(feats_df = glb_feats_df, entity_df =
## glb_entity_df, : Identified reimbursement2008 as highly correlated with
## other features
## [1] "checking correlations for features:"
## [1] "bucket2008" "ihd" "diabetes" "kidney"
## [5] "heart.failure" "copd" "depression" "alzheimers"
## [9] "arthritis" "cancer" "osteoporosis" "stroke"
## [13] "age" ".rnorm"
## bucket2008 ihd diabetes kidney
## bucket2008 1.000000000 0.457745311 0.441695494 0.556443060
## ihd 0.457745311 1.000000000 0.514681463 0.379348673
## diabetes 0.441695494 0.514681463 1.000000000 0.410508010
## kidney 0.556443060 0.379348673 0.410508010 1.000000000
## heart.failure 0.483790416 0.466853617 0.447713361 0.418535903
## copd 0.479911725 0.341479503 0.337436848 0.362145211
## depression 0.342770130 0.323337325 0.347401660 0.274868719
## alzheimers 0.375421748 0.326534346 0.341021686 0.304331472
## arthritis 0.330250474 0.311822976 0.323405716 0.250350192
## cancer 0.295879956 0.192569561 0.182463340 0.187306219
## osteoporosis 0.234597479 0.265948840 0.273431836 0.192706719
## stroke 0.319686629 0.193566819 0.191335431 0.229542821
## age 0.065957565 0.072806906 0.067414459 0.056523917
## .rnorm -0.005471367 -0.002117255 -0.001223323 -0.006698442
## heart.failure copd depression alzheimers
## bucket2008 0.48379042 0.479911725 0.342770130 0.375421748
## ihd 0.46685362 0.341479503 0.323337325 0.326534346
## diabetes 0.44771336 0.337436848 0.347401660 0.341021686
## kidney 0.41853590 0.362145211 0.274868719 0.304331472
## heart.failure 1.00000000 0.376619956 0.303108016 0.331637621
## copd 0.37661996 1.000000000 0.262091958 0.282316055
## depression 0.30310802 0.262091958 1.000000000 0.289283692
## alzheimers 0.33163762 0.282316055 0.289283692 1.000000000
## arthritis 0.26577236 0.218711881 0.231412945 0.205787238
## cancer 0.17399979 0.173576229 0.120859944 0.136229360
## osteoporosis 0.22170825 0.184441404 0.193200212 0.189830404
## stroke 0.22049646 0.206759478 0.155228820 0.217944953
## age 0.06758002 0.044475985 0.013183881 0.044509085
## .rnorm -0.00413497 -0.003209717 -0.004546531 -0.002242699
## arthritis cancer osteoporosis stroke
## bucket2008 0.330250474 0.295879956 0.2345974786 0.319686629
## ihd 0.311822976 0.192569561 0.2659488405 0.193566819
## diabetes 0.323405716 0.182463340 0.2734318364 0.191335431
## kidney 0.250350192 0.187306219 0.1927067194 0.229542821
## heart.failure 0.265772357 0.173999790 0.2217082477 0.220496456
## copd 0.218711881 0.173576229 0.1844414035 0.206759478
## depression 0.231412945 0.120859944 0.1932002120 0.155228820
## alzheimers 0.205787238 0.136229360 0.1898304044 0.217944953
## arthritis 1.000000000 0.116224739 0.2188952357 0.117518638
## cancer 0.116224739 1.000000000 0.1145554395 0.100858117
## osteoporosis 0.218895236 0.114555439 1.0000000000 0.100034864
## stroke 0.117518638 0.100858117 0.1000348636 1.000000000
## age 0.042003908 0.037706744 0.0400590867 0.036510287
## .rnorm -0.003328174 -0.001080007 -0.0007172342 -0.003483414
## age .rnorm
## bucket2008 0.0659575650 -0.0054713667
## ihd 0.0728069059 -0.0021172555
## diabetes 0.0674144586 -0.0012233233
## kidney 0.0565239171 -0.0066984421
## heart.failure 0.0675800164 -0.0041349700
## copd 0.0444759850 -0.0032097166
## depression 0.0131838813 -0.0045465313
## alzheimers 0.0445090852 -0.0022426985
## arthritis 0.0420039076 -0.0033281745
## cancer 0.0377067437 -0.0010800072
## osteoporosis 0.0400590867 -0.0007172342
## stroke 0.0365102875 -0.0034834136
## age 1.0000000000 0.0007606978
## .rnorm 0.0007606978 1.0000000000
## bucket2008 ihd diabetes kidney
## bucket2008 0.000000000 0.457745311 0.441695494 0.556443060
## ihd 0.457745311 0.000000000 0.514681463 0.379348673
## diabetes 0.441695494 0.514681463 0.000000000 0.410508010
## kidney 0.556443060 0.379348673 0.410508010 0.000000000
## heart.failure 0.483790416 0.466853617 0.447713361 0.418535903
## copd 0.479911725 0.341479503 0.337436848 0.362145211
## depression 0.342770130 0.323337325 0.347401660 0.274868719
## alzheimers 0.375421748 0.326534346 0.341021686 0.304331472
## arthritis 0.330250474 0.311822976 0.323405716 0.250350192
## cancer 0.295879956 0.192569561 0.182463340 0.187306219
## osteoporosis 0.234597479 0.265948840 0.273431836 0.192706719
## stroke 0.319686629 0.193566819 0.191335431 0.229542821
## age 0.065957565 0.072806906 0.067414459 0.056523917
## .rnorm 0.005471367 0.002117255 0.001223323 0.006698442
## heart.failure copd depression alzheimers
## bucket2008 0.48379042 0.479911725 0.342770130 0.375421748
## ihd 0.46685362 0.341479503 0.323337325 0.326534346
## diabetes 0.44771336 0.337436848 0.347401660 0.341021686
## kidney 0.41853590 0.362145211 0.274868719 0.304331472
## heart.failure 0.00000000 0.376619956 0.303108016 0.331637621
## copd 0.37661996 0.000000000 0.262091958 0.282316055
## depression 0.30310802 0.262091958 0.000000000 0.289283692
## alzheimers 0.33163762 0.282316055 0.289283692 0.000000000
## arthritis 0.26577236 0.218711881 0.231412945 0.205787238
## cancer 0.17399979 0.173576229 0.120859944 0.136229360
## osteoporosis 0.22170825 0.184441404 0.193200212 0.189830404
## stroke 0.22049646 0.206759478 0.155228820 0.217944953
## age 0.06758002 0.044475985 0.013183881 0.044509085
## .rnorm 0.00413497 0.003209717 0.004546531 0.002242699
## arthritis cancer osteoporosis stroke
## bucket2008 0.330250474 0.295879956 0.2345974786 0.319686629
## ihd 0.311822976 0.192569561 0.2659488405 0.193566819
## diabetes 0.323405716 0.182463340 0.2734318364 0.191335431
## kidney 0.250350192 0.187306219 0.1927067194 0.229542821
## heart.failure 0.265772357 0.173999790 0.2217082477 0.220496456
## copd 0.218711881 0.173576229 0.1844414035 0.206759478
## depression 0.231412945 0.120859944 0.1932002120 0.155228820
## alzheimers 0.205787238 0.136229360 0.1898304044 0.217944953
## arthritis 0.000000000 0.116224739 0.2188952357 0.117518638
## cancer 0.116224739 0.000000000 0.1145554395 0.100858117
## osteoporosis 0.218895236 0.114555439 0.0000000000 0.100034864
## stroke 0.117518638 0.100858117 0.1000348636 0.000000000
## age 0.042003908 0.037706744 0.0400590867 0.036510287
## .rnorm 0.003328174 0.001080007 0.0007172342 0.003483414
## age .rnorm
## bucket2008 0.0659575650 0.0054713667
## ihd 0.0728069059 0.0021172555
## diabetes 0.0674144586 0.0012233233
## kidney 0.0565239171 0.0066984421
## heart.failure 0.0675800164 0.0041349700
## copd 0.0444759850 0.0032097166
## depression 0.0131838813 0.0045465313
## alzheimers 0.0445090852 0.0022426985
## arthritis 0.0420039076 0.0033281745
## cancer 0.0377067437 0.0010800072
## osteoporosis 0.0400590867 0.0007172342
## stroke 0.0365102875 0.0034834136
## age 0.0000000000 0.0007606978
## .rnorm 0.0007606978 0.0000000000
## id cor.y exclude.as.feat
## bucket2009 bucket2009 1.0000000000 1
## reimbursement2009 reimbursement2009 0.8581631864 1
## bucket2008 bucket2008 0.4518935763 0
## bucket2008.fctr bucket2008.fctr 0.4518935763 1
## ihd ihd 0.3905905884 0
## diabetes diabetes 0.3904719536 0
## reimbursement2008 reimbursement2008 0.3756473063 0
## kidney kidney 0.3683780944 0
## heart.failure heart.failure 0.3647689526 0
## copd copd 0.3108325355 0
## depression depression 0.2835366153 0
## alzheimers alzheimers 0.2741643394 0
## arthritis arthritis 0.2717113526 0
## cancer cancer 0.2100892954 0
## osteoporosis osteoporosis 0.2076745377 0
## stroke stroke 0.1846626746 0
## age age 0.0495694151 0
## .rnorm .rnorm -0.0007970401 0
## cor.y.abs cor.low
## bucket2009 1.0000000000 0
## reimbursement2009 0.8581631864 0
## bucket2008 0.4518935763 1
## bucket2008.fctr 0.4518935763 0
## ihd 0.3905905884 1
## diabetes 0.3904719536 1
## reimbursement2008 0.3756473063 0
## kidney 0.3683780944 1
## heart.failure 0.3647689526 1
## copd 0.3108325355 1
## depression 0.2835366153 1
## alzheimers 0.2741643394 1
## arthritis 0.2717113526 1
## cancer 0.2100892954 1
## osteoporosis 0.2076745377 1
## stroke 0.1846626746 1
## age 0.0495694151 1
## .rnorm 0.0007970401 1glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="fit.models",
chunk_step_major=max(glb_script_df$chunk_step_major)+1,
chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor
## elapsed7 remove_correlated_features 4 1
## elapsed8 fit.models 5 0
## elapsed
## elapsed7 29.537
## elapsed8 71.848Step 5: fit models
max_cor_y_x_var <- subset(glb_feats_df, cor.low == 1)[1, "id"]
if (!is.null(glb_Baseline_mdl_var)) {
if ((max_cor_y_x_var != glb_Baseline_mdl_var) &
(glb_feats_df[max_cor_y_x_var, "cor.y.abs"] >
glb_feats_df[glb_Baseline_mdl_var, "cor.y.abs"]))
stop(max_cor_y_x_var, " has a lower correlation with ", glb_rsp_var,
" than the Baseline var: ", glb_Baseline_mdl_var)
}
# Regression:
if (glb_is_regression) {
# Linear:
myfit_mdl_fn <- myfit_mdl_lm
}
# Classification:
if (glb_is_classification) myfit_mdl_fn <- myfit_mdl_classification
glb_is_binomial <- (length(unique(glb_entity_df[, glb_rsp_var])) <= 2)
# Any models that have tuning parameters has "better" results with cross-validation
# & "different" results for different outcome metrics
# Baseline
if (!is.null(glb_Baseline_mdl_var)) {
# lm_mdl <- lm(reformulate(glb_Baseline_mdl_var,
# response="bucket2009"), data=glb_entity_df)
# print(summary(lm_mdl))
# plot(lm_mdl, ask=FALSE)
# ret_lst <- myfit_mdl_fn(model_id="Baseline",
# model_method=ifelse(glb_is_regression, "lm",
# ifelse(glb_is_binomial, "glm", "rpart")),
# indep_vars_vctr=glb_Baseline_mdl_var,
# rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
# fit_df=glb_entity_df, OOB_df=glb_newent_df,
# n_cv_folds=0, tune_models_df=NULL,
# model_loss_mtrx=glb_model_metric_terms,
# model_summaryFunction=glb_model_metric_smmry,
# model_metric=glb_model_metric,
# model_metric_maximize=glb_model_metric_maximize)
ret_lst <- myfit_mdl_fn(model_id="Baseline", model_method="mybaseln_classfr",
indep_vars_vctr=glb_Baseline_mdl_var,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df)
}## Loading required package: caret
## Loading required package: lattice
##
## Attaching package: 'caret'
##
## The following object is masked from 'package:survival':
##
## cluster## [1] "fitting model: Baseline.mybaseln_classfr"
## [1] " indep_vars: bucket2008.fctr, .rnorm"## Warning in if (model_method == "rf") methodControl <- "oob": the condition
## has length > 1 and only the first element will be used## Fitting parameter = none on full training set
## [1] "in Baseline.Classifier$fit"
## [1] "class(x):"
## [1] "matrix"
## [1] "dimnames(x)[[2]]:"
## [1] "bucket2008.fctrB2" "bucket2008.fctrB3" "bucket2008.fctrB4"
## [4] "bucket2008.fctrB5" ".rnorm"
## [1] "length(x):"
## [1] 1374015
## [1] "head(x):"
## bucket2008.fctrB2 bucket2008.fctrB3 bucket2008.fctrB4 bucket2008.fctrB5
## 1 0 0 0 0
## 2 0 0 0 0
## 4 0 0 0 0
## 7 0 0 0 0
## 11 0 0 0 0
## 13 0 0 0 0
## .rnorm
## 1 -0.9248001
## 2 0.1533902
## 4 -0.9917587
## 7 -0.5160354
## 11 -1.0376029
## 13 -0.3801495
## [1] "class(y):"
## [1] "factor"
## [1] "length(y):"
## [1] 274803
## [1] "head(y):"
## 1 2 4 7 11 13
## B1 B1 B1 B1 B1 B1
## Levels: B1 B2 B3 B4 B5
## Length Class Mode
## x_names 4 -none- character
## x_vals 5 -none- character
## xNames 5 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 5 -none- character
## [1] "in Baseline.Classifier$predict"
## [1] "class(newdata):"
## [1] "matrix"
## [1] "head(newdata):"
## bucket2008.fctrB2 bucket2008.fctrB3 bucket2008.fctrB4 bucket2008.fctrB5
## 1 0 0 0 0
## 2 0 0 0 0
## 4 0 0 0 0
## 7 0 0 0 0
## 11 0 0 0 0
## 13 0 0 0 0
## .rnorm
## 1 -0.9248001
## 2 0.1533902
## 4 -0.9917587
## 7 -0.5160354
## 11 -1.0376029
## 13 -0.3801495
## [1] "x_names: "
## [1] "bucket2008.fctrB2" "bucket2008.fctrB3" "bucket2008.fctrB4"
## [4] "bucket2008.fctrB5"
## [1] "x_vals: "
## [1] "B1" "B2" "B3" "B4" "B5"
## [1] "length(y):"
## [1] 274803
## [1] "head(y):"
## [1] B1 B1 B1 B1 B1 B1
## Levels: B1 B2 B3 B4 B5## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 164967 11747 5214 2275 263
## B2 24001 16172 6877 4367 842
## B3 10679 6848 4004 2552 503
## B4 4020 2835 2081 2399 571
## B5 410 300 266 469 141
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.829729e-01 NA 6.812294e-01 6.847125e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 1.601032e-39 0.000000e+00
## [1] "in Baseline.Classifier$predict"
## [1] "class(newdata):"
## [1] "matrix"
## [1] "head(newdata):"
## bucket2008.fctrB2 bucket2008.fctrB3 bucket2008.fctrB4 bucket2008.fctrB5
## 3 0 0 0 0
## 5 0 0 0 0
## 6 0 0 0 0
## 8 0 0 0 0
## 9 0 0 0 0
## 10 0 0 0 0
## .rnorm
## 3 0.7183313438
## 5 0.0008759006
## 6 -1.4428971189
## 8 -1.9680342619
## 9 -3.0026827087
## 10 -1.1723168041
## [1] "x_names: "
## [1] "bucket2008.fctrB2" "bucket2008.fctrB3" "bucket2008.fctrB4"
## [4] "bucket2008.fctrB5"
## [1] "x_vals: "
## [1] "B1" "B2" "B3" "B4" "B5"
## [1] "length(y):"
## [1] 183202
## [1] "head(y):"
## [1] B1 B1 B1 B1 B1 B1
## Levels: B1 B2 B3 B4 B5## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110138 7787 3427 1452 174
## B2 16000 10721 4629 2931 559
## B3 7006 4629 2774 1621 360
## B4 2688 1943 1415 1539 352
## B5 293 191 160 309 104
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.838135e-01 NA 6.816786e-01 6.859425e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 9.943570e-31 0.000000e+00
## model_id model_method feats
## 1 Baseline.mybaseln_classfr mybaseln_classfr bucket2008.fctr, .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 0 0.626 0.006
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.6829729 0.6812294 0.6847125
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.6838135 0.6816786
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit
## 1 0.6859425 NA 0# Most Frequent Outcome "MFO" model: mean(y) for regression
# Not using caret's nullModel since model stats not avl
# Cannot use rpart for multinomial classification since it predicts non-MFO
ret_lst <- myfit_mdl_fn(model_id="MFO", model_method="myMFO_classfr",
indep_vars_vctr=".rnorm",
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df)## [1] "fitting model: MFO.myMFO_classfr"
## [1] " indep_vars: .rnorm"## Warning in if (model_method == "rf") methodControl <- "oob": the condition
## has length > 1 and only the first element will be used## Fitting parameter = none on full training set
## [1] "in MFO.Classifier$fit"
## [1] "unique.vals:"
## [1] B1 B2 B3 B4 B5
## Levels: B1 B2 B3 B4 B5
## [1] "unique.prob:"
## y
## B1 B2 B3 B4 B5
## 0.671266325 0.190168957 0.089467728 0.043325582 0.005771407
## [1] "MFO.val:"
## [1] "B1"
## Length Class Mode
## unique.vals 5 factor numeric
## unique.prob 5 -none- numeric
## MFO.val 1 -none- character
## xNames 1 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 5 -none- character
## [1] "in MFO.Classifier$predict"## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 184466 0 0 0 0
## B2 52259 0 0 0 0
## B3 24586 0 0 0 0
## B4 11906 0 0 0 0
## B5 1586 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712663 NA 0.6695064 0.6730227 0.6712663
## AccuracyPValue McnemarPValue
## 0.5009025 NaN
## [1] "in MFO.Classifier$predict"## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 122978 0 0 0 0
## B2 34840 0 0 0 0
## B3 16390 0 0 0 0
## B4 7937 0 0 0 0
## B5 1057 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712700 NA 0.6691135 0.6734210 0.6712700
## AccuracyPValue McnemarPValue
## 0.5011054 NaN
## model_id model_method feats max.nTuningRuns
## 1 MFO.myMFO_classfr myMFO_classfr .rnorm 0
## min.elapsedtime.everything min.elapsedtime.final max.Accuracy.fit
## 1 0.515 0.038 0.6712663
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.6695064 0.6730227 NA
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.67127 0.6691135 0.673421
## max.Kappa.OOB min.SSE.fit
## 1 NA 0# "random" model - only for classification; none needed for regression since it is same as MFO
ret_lst <- myfit_mdl_fn(model_id="Random", model_method="myrandom_classfr",
indep_vars_vctr=".rnorm",
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df)## [1] "fitting model: Random.myrandom_classfr"
## [1] " indep_vars: .rnorm"## Warning in if (model_method == "rf") methodControl <- "oob": the condition
## has length > 1 and only the first element will be used## Fitting parameter = none on full training set
## Length Class Mode
## unique.vals 5 factor numeric
## unique.prob 5 table numeric
## xNames 1 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 5 -none- character## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 123841 35132 16435 7984 1074
## B2 35132 9946 4671 2224 286
## B3 16483 4699 2202 1081 121
## B4 8003 2259 1085 503 56
## B5 1056 325 127 68 10
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.4967267 NA 0.4948556 0.4985980 0.6712663
## AccuracyPValue McnemarPValue
## 1.0000000 0.9272432## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 82207 23808 10990 5253 720
## B2 23379 6626 3141 1506 188
## B3 11037 3119 1460 683 91
## B4 5329 1497 705 364 42
## B5 707 204 97 43 6
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.4948800 NA 0.4925879 0.4971722 0.6712700
## AccuracyPValue McnemarPValue
## 1.0000000 0.8174342
## model_id model_method feats max.nTuningRuns
## 1 Random.myrandom_classfr myrandom_classfr .rnorm 0
## min.elapsedtime.everything min.elapsedtime.final max.Accuracy.fit
## 1 0.375 0.035 0.4967267
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.4948556 0.498598 NA
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.49488 0.4925879 0.4971722
## max.Kappa.OOB min.SSE.fit
## 1 NA 0# Max.cor.Y
ret_lst <- myfit_mdl_fn(model_id="Max.cor.Y.cv.0",
model_method=ifelse(glb_is_regression, "lm",
ifelse(glb_is_binomial, "glm", "rpart")),
indep_vars_vctr=max_cor_y_x_var,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df)## [1] "fitting model: Max.cor.Y.cv.0.rpart"
## [1] " indep_vars: bucket2008"## Loading required package: rpart## Fitting cp = 0.097 on full training set## Loading required package: rpart.plot## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 0.09695916 0 1
##
## Node number 1: 274803 observations
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 184466 0 0 0 0
## B2 52259 0 0 0 0
## B3 24586 0 0 0 0
## B4 11906 0 0 0 0
## B5 1586 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712663 NA 0.6695064 0.6730227 0.6712663
## AccuracyPValue McnemarPValue
## 0.5009025 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## Prediction
## Reference B1 B2 B3 B4 B5
## B1 122978 0 0 0 0
## B2 34840 0 0 0 0
## B3 16390 0 0 0 0
## B4 7937 0 0 0 0
## B5 1057 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712700 NA 0.6691135 0.6734210 0.6712700
## AccuracyPValue McnemarPValue
## 0.5011054 NaN
## model_id model_method feats max.nTuningRuns
## 1 Max.cor.Y.cv.0.rpart rpart bucket2008 0
## min.elapsedtime.everything min.elapsedtime.final max.Accuracy.fit
## 1 4.811 3.43 0.6712663
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.6695064 0.6730227 NA
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.67127 0.6691135 0.673421
## max.Kappa.OOB min.SSE.fit
## 1 NA 0ret_lst <- myfit_mdl_fn(model_id="Max.cor.Y.cv.G",
model_method=ifelse(glb_is_regression, "lm",
ifelse(glb_is_binomial, "glm", "rpart")),
indep_vars_vctr=max_cor_y_x_var,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=NULL)## [1] "fitting model: Max.cor.Y.cv.G.rpart"
## [1] " indep_vars: bucket2008"
## + Fold1: cp=0
## - Fold1: cp=0
## + Fold2: cp=0
## - Fold2: cp=0
## + Fold3: cp=0
## - Fold3: cp=0
## + Fold4: cp=0
## - Fold4: cp=0
## + Fold5: cp=0
## - Fold5: cp=0
## Aggregating results
## Selecting tuning parameters
## Fitting cp = 0.0485 on full training set
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 0.09695916 0 1.0000000
## 2 0.00000000 1 0.9030408
##
## Variable importance
## bucket2008
## 100
##
## Node number 1: 274803 observations, complexity param=0.09695916
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (204077 obs) right son=3 (70726 obs)
## Primary splits:
## bucket2008 < 1.5 to the left, improve=20624.7, (0 missing)
##
## Node number 2: 204077 observations
## predicted class=B1 expected loss=0.1916434 P(node) =0.7426302
## class counts: 164967 24001 10679 4020 410
## probabilities: 0.808 0.118 0.052 0.020 0.002
##
## Node number 3: 70726 observations
## predicted class=B2 expected loss=0.6004581 P(node) =0.2573698
## class counts: 19499 28258 13907 7886 1176
## probabilities: 0.276 0.400 0.197 0.112 0.017
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) bucket2008< 1.5 204077 39110 B1 (0.81 0.12 0.052 0.02 0.002) *
## 3) bucket2008>=1.5 70726 42468 B2 (0.28 0.4 0.2 0.11 0.017) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 164967 19499 0 0 0
## B2 24001 28258 0 0 0
## B3 10679 13907 0 0 0
## B4 4020 7886 0 0 0
## B5 410 1176 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.031401e-01 NA 7.014279e-01 7.048479e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 3.041172e-282 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110138 12840 0 0 0
## B2 16000 18840 0 0 0
## B3 7006 9384 0 0 0
## B4 2688 5249 0 0 0
## B5 293 764 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.040207e-01 NA 7.019245e-01 7.061105e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 1.813355e-199 NaN
## model_id model_method feats max.nTuningRuns
## 1 Max.cor.Y.cv.G.rpart rpart bucket2008 3
## min.elapsedtime.everything min.elapsedtime.final max.Accuracy.fit
## 1 20.801 3.333 0.7031401
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.7014279 0.7048479 0.3440289
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7040207 0.7019245 0.7061105
## max.Kappa.OOB min.SSE.fit max.AccuracySD.fit max.KappaSD.fit
## 1 NA 0 0.001665788 0.004508795# Interactions.High.cor.Y
if (nrow(int_feats_df <- subset(glb_feats_df, (cor.low == 0) &
(exclude.as.feat == 0))) > 0) {
# Only glm handles interaction terms (checked that rpart does not)
# This does not work - why ???
# indep_vars_vctr <- ifelse(glb_is_binomial,
# c(max_cor_y_x_var, paste(max_cor_y_x_var,
# subset(glb_feats_df, is.na(cor.low))[, "id"], sep=":")),
# union(max_cor_y_x_var, subset(glb_feats_df, is.na(cor.low))[, "id"]))
if (glb_is_regression | glb_is_binomial) {
indep_vars_vctr <-
c(max_cor_y_x_var, paste(max_cor_y_x_var, int_feats_df[, "id"], sep=":"))
} else { indep_vars_vctr <- union(max_cor_y_x_var, int_feats_df[, "id"]) }
ret_lst <- myfit_mdl_fn(model_id="Interact.High.cor.y",
model_method=ifelse(glb_is_regression, "lm",
ifelse(glb_is_binomial, "glm", "rpart")),
indep_vars_vctr,
glb_rsp_var, glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=NULL)
} ## [1] "fitting model: Interact.High.cor.y.rpart"
## [1] " indep_vars: bucket2008, reimbursement2008"
## + Fold1: cp=4.428e-05
## - Fold1: cp=4.428e-05
## + Fold2: cp=4.428e-05
## - Fold2: cp=4.428e-05
## + Fold3: cp=4.428e-05
## - Fold3: cp=4.428e-05
## + Fold4: cp=4.428e-05
## - Fold4: cp=4.428e-05
## + Fold5: cp=4.428e-05
## - Fold5: cp=4.428e-05
## Aggregating results
## Selecting tuning parameters
## Fitting cp = 4.67e-05 on full training set
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 4.908841e-02 0 1.0000000
## 2 4.665072e-05 2 0.9018232
##
## Variable importance
## reimbursement2008 bucket2008
## 60 40
##
## Node number 1: 274803 observations, complexity param=0.04908841
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (165987 obs) right son=3 (108816 obs)
## Primary splits:
## reimbursement2008 < 1565 to the left, improve=24395.14, (0 missing)
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.861, adj=0.65, (0 split)
##
## Node number 2: 165987 observations
## predicted class=B1 expected loss=0.1261424 P(node) =0.6040218
## class counts: 145049 12284 6102 2315 237
## probabilities: 0.874 0.074 0.037 0.014 0.001
##
## Node number 3: 108816 observations, complexity param=0.04908841
## predicted class=B2 expected loss=0.6326367 P(node) =0.3959782
## class counts: 39417 39975 18484 9591 1349
## probabilities: 0.362 0.367 0.170 0.088 0.012
## left son=6 (39298 obs) right son=7 (69518 obs)
## Primary splits:
## reimbursement2008 < 3065 to the left, improve=2010.308, (0 missing)
## bucket2008 < 1.5 to the left, improve=1980.977, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.989, adj=0.969, (0 split)
##
## Node number 6: 39298 observations
## predicted class=B1 expected loss=0.4797445 P(node) =0.1430043
## class counts: 20445 12134 4756 1782 181
## probabilities: 0.520 0.309 0.121 0.045 0.005
##
## Node number 7: 69518 observations
## predicted class=B2 expected loss=0.5995138 P(node) =0.2529739
## class counts: 18972 27841 13728 7809 1168
## probabilities: 0.273 0.400 0.197 0.112 0.017
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) reimbursement2008< 1565 165987 20938 B1 (0.87 0.074 0.037 0.014 0.0014) *
## 3) reimbursement2008>=1565 108816 68841 B2 (0.36 0.37 0.17 0.088 0.012)
## 6) reimbursement2008< 3065 39298 18853 B1 (0.52 0.31 0.12 0.045 0.0046) *
## 7) reimbursement2008>=3065 69518 41677 B2 (0.27 0.4 0.2 0.11 0.017) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 165494 18972 0 0 0
## B2 24418 27841 0 0 0
## B3 10858 13728 0 0 0
## B4 4097 7809 0 0 0
## B5 418 1168 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.035404e-01 NA 7.018289e-01 7.052475e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 2.207353e-289 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110452 12526 0 0 0
## B2 16322 18518 0 0 0
## B3 7105 9285 0 0 0
## B4 2740 5197 0 0 0
## B5 299 758 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.039770e-01 NA 7.018807e-01 7.060669e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 6.148674e-199 NaN
## model_id model_method feats
## 1 Interact.High.cor.y.rpart rpart bucket2008, reimbursement2008
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 3 31.301 5.05
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.70298 0.7018289 0.7052475
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 0.3368565 0.703977 0.7018807
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.7060669 NA 0 0.001690032
## max.KappaSD.fit
## 1 0.006361514# Low.cor.X
ret_lst <- myfit_mdl_fn(model_id="Low.cor.X",
model_method=ifelse(glb_is_regression, "lm",
ifelse(glb_is_binomial, "glm", "rpart")),
indep_vars_vctr=subset(glb_feats_df, cor.low == 1)[, "id"],
glb_rsp_var, glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=NULL)## [1] "fitting model: Low.cor.X.rpart"
## [1] " indep_vars: bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age"
## + Fold1: cp=0.004317
## - Fold1: cp=0.004317
## + Fold2: cp=0.004317
## - Fold2: cp=0.004317
## + Fold3: cp=0.004317
## - Fold3: cp=0.004317
## + Fold4: cp=0.004317
## - Fold4: cp=0.004317
## + Fold5: cp=0.004317
## - Fold5: cp=0.004317
## Aggregating results
## Selecting tuning parameters
## Fitting cp = 0.00432 on full training set## Warning in myfit_mdl_fn(model_id = "Low.cor.X", model_method =
## ifelse(glb_is_regression, : model's bestTune found at an extreme of
## tuneGrid for parameter: cp
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 0.096959164 0 1.0000000
## 2 0.015010461 1 0.9030408
## 3 0.004317168 2 0.8880304
##
## Variable importance
## bucket2008 kidney copd heart.failure arthritis
## 49 15 12 9 7
## alzheimers diabetes
## 6 2
##
## Node number 1: 274803 observations, complexity param=0.09695916
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (204077 obs) right son=3 (70726 obs)
## Primary splits:
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## ihd < 0.5 to the left, improve=16291.74, (0 missing)
## diabetes < 0.5 to the left, improve=16041.26, (0 missing)
## heart.failure < 0.5 to the left, improve=12498.16, (0 missing)
## kidney < 0.5 to the left, improve=10965.93, (0 missing)
## Surrogate splits:
## kidney < 0.5 to the left, agree=0.823, adj=0.311, (0 split)
## copd < 0.5 to the left, agree=0.806, adj=0.246, (0 split)
## heart.failure < 0.5 to the left, agree=0.791, adj=0.187, (0 split)
## arthritis < 0.5 to the left, agree=0.782, adj=0.153, (0 split)
## alzheimers < 0.5 to the left, agree=0.774, adj=0.120, (0 split)
##
## Node number 2: 204077 observations
## predicted class=B1 expected loss=0.1916434 P(node) =0.7426302
## class counts: 164967 24001 10679 4020 410
## probabilities: 0.808 0.118 0.052 0.020 0.002
##
## Node number 3: 70726 observations, complexity param=0.01501046
## predicted class=B2 expected loss=0.6004581 P(node) =0.2573698
## class counts: 19499 28258 13907 7886 1176
## probabilities: 0.276 0.400 0.197 0.112 0.017
## left son=6 (16128 obs) right son=7 (54598 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=662.2867, (0 missing)
## kidney < 0.5 to the left, improve=620.6834, (0 missing)
## arthritis < 0.5 to the left, improve=517.1168, (0 missing)
## ihd < 0.5 to the left, improve=443.6660, (0 missing)
## heart.failure < 0.5 to the left, improve=397.1855, (0 missing)
##
## Node number 6: 16128 observations
## predicted class=B1 expected loss=0.5680184 P(node) =0.05868932
## class counts: 6967 5611 2420 1017 113
## probabilities: 0.432 0.348 0.150 0.063 0.007
##
## Node number 7: 54598 observations
## predicted class=B2 expected loss=0.5852046 P(node) =0.1986805
## class counts: 12532 22647 11487 6869 1063
## probabilities: 0.230 0.415 0.210 0.126 0.019
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) bucket2008< 1.5 204077 39110 B1 (0.81 0.12 0.052 0.02 0.002) *
## 3) bucket2008>=1.5 70726 42468 B2 (0.28 0.4 0.2 0.11 0.017)
## 6) diabetes< 0.5 16128 9161 B1 (0.43 0.35 0.15 0.063 0.007) *
## 7) diabetes>=0.5 54598 31951 B2 (0.23 0.41 0.21 0.13 0.019) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 171934 12532 0 0 0
## B2 29612 22647 0 0 0
## B3 13099 11487 0 0 0
## B4 5037 6869 0 0 0
## B5 523 1063 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7080745 NA 0.7063706 0.7097740 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114716 8262 0 0 0
## B2 19896 14944 0 0 0
## B3 8672 7718 0 0 0
## B4 3366 4571 0 0 0
## B5 379 678 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.077434e-01 NA 7.056548e-01 7.098254e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 2.343725e-247 NaN
## model_id model_method
## 1 Low.cor.X.rpart rpart
## feats
## 1 bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 3 101.182 13.937
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.7086058 0.7063706 0.709774
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 0.3161651 0.7077434 0.7056548
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.7098254 NA 0 0.0009742811
## max.KappaSD.fit
## 1 0.004868262# User specified
for (method in glb_models_method_vctr) {
print(sprintf("iterating over method:%s", method))
# All X that is not user excluded
indep_vars_vctr <- setdiff(names(glb_entity_df),
union(glb_rsp_var, glb_exclude_vars_as_features))
# easier to exclude features
# indep_vars_vctr <- setdiff(names(glb_entity_df),
# union(union(glb_rsp_var, glb_exclude_vars_as_features),
# c("<feat1_name>", "<feat2_name>")))
# easier to include features
# indep_vars_vctr <- c("<feat1_name>", "<feat2_name>")
# User specified bivariate models
# indep_vars_vctr_lst <- list()
# for (feat in setdiff(names(glb_entity_df),
# union(glb_rsp_var, glb_exclude_vars_as_features)))
# indep_vars_vctr_lst[["feat"]] <- feat
# User specified combinatorial models
# indep_vars_vctr_lst <- list()
# combn_mtrx <- combn(c("<feat1_name>", "<feat2_name>", "<featn_name>"),
# <num_feats_to_choose>)
# for (combn_ix in 1:ncol(combn_mtrx))
# #print(combn_mtrx[, combn_ix])
# indep_vars_vctr_lst[[combn_ix]] <- combn_mtrx[, combn_ix]
# glb_sel_mdl <- glb_sel_wlm_mdl <- ret_lst[["model"]]
# rpart_sel_wlm_mdl <- rpart(reformulate(indep_vars_vctr, response=glb_rsp_var),
# data=glb_entity_df, method="class",
# control=rpart.control(cp=glb_sel_wlm_mdl$bestTune$cp),
# parms=list(loss=glb_model_metric_terms))
# print("rpart_sel_wlm_mdl"); prp(rpart_sel_wlm_mdl)
#
model_id_pfx <- "All.X";
ret_lst <- myfit_mdl_fn(model_id=paste0(model_id_pfx, ".lser.no.cp.opt"), model_method=method,
indep_vars_vctr=indep_vars_vctr,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=NULL)
if (method == "rpart")
ret_lst <- myfit_mdl_fn(model_id=paste0(model_id_pfx, ".lser.no.cp.4015"), model_method=method,
indep_vars_vctr=indep_vars_vctr,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=glb_tune_models_df)
# rf is hard-coded in caret to recognize only Accuracy / Kappa evaluation metrics
# only for OOB in trainControl ?
ret_lst <- myfit_mdl_fn(model_id=paste0(model_id_pfx, ".lser.ys.cp.opt"), model_method=method,
indep_vars_vctr=indep_vars_vctr,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=NULL,
model_loss_mtrx=glb_model_metric_terms,
model_summaryFunction=glb_model_metric_smmry,
model_metric=glb_model_metric,
model_metric_maximize=glb_model_metric_maximize)
if (method == "rpart")
ret_lst <- myfit_mdl_fn(model_id=paste0(model_id_pfx, ".lser.ys.cp.4015"), model_method=method,
indep_vars_vctr=indep_vars_vctr,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df, OOB_df=glb_newent_df,
n_cv_folds=glb_n_cv_folds, tune_models_df=glb_tune_models_df,
model_loss_mtrx=glb_model_metric_terms,
model_summaryFunction=glb_model_metric_smmry,
model_metric=glb_model_metric,
model_metric_maximize=glb_model_metric_maximize)
}## [1] "iterating over method:rpart"
## [1] "fitting model: All.X.lser.no.cp.opt.rpart"
## [1] " indep_vars: age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008"
## + Fold1: cp=0.00435
## - Fold1: cp=0.00435
## + Fold2: cp=0.00435
## - Fold2: cp=0.00435
## + Fold3: cp=0.00435
## - Fold3: cp=0.00435
## + Fold4: cp=0.00435
## - Fold4: cp=0.00435
## + Fold5: cp=0.00435
## - Fold5: cp=0.00435
## Aggregating results
## Selecting tuning parameters
## Fitting cp = 0.00435 on full training set## Warning in myfit_mdl_fn(model_id = paste0(model_id_pfx,
## ".lser.no.cp.opt"), : model's bestTune found at an extreme of tuneGrid for
## parameter: cp
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 0.049088413 0 1.0000000
## 2 0.013958843 2 0.9018232
## 3 0.004350377 3 0.8878643
##
## Variable importance
## reimbursement2008 bucket2008 diabetes ihd
## 31 21 14 14
## heart.failure kidney
## 11 9
##
## Node number 1: 274803 observations, complexity param=0.04908841
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (165987 obs) right son=3 (108816 obs)
## Primary splits:
## reimbursement2008 < 1565 to the left, improve=24395.14, (0 missing)
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## ihd < 0.5 to the left, improve=16291.74, (0 missing)
## diabetes < 0.5 to the left, improve=16041.26, (0 missing)
## heart.failure < 0.5 to the left, improve=12498.16, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.861, adj=0.650, (0 split)
## ihd < 0.5 to the left, agree=0.792, adj=0.474, (0 split)
## diabetes < 0.5 to the left, agree=0.785, adj=0.456, (0 split)
## heart.failure < 0.5 to the left, agree=0.762, adj=0.399, (0 split)
## kidney < 0.5 to the left, agree=0.731, adj=0.321, (0 split)
##
## Node number 2: 165987 observations
## predicted class=B1 expected loss=0.1261424 P(node) =0.6040218
## class counts: 145049 12284 6102 2315 237
## probabilities: 0.874 0.074 0.037 0.014 0.001
##
## Node number 3: 108816 observations, complexity param=0.04908841
## predicted class=B2 expected loss=0.6326367 P(node) =0.3959782
## class counts: 39417 39975 18484 9591 1349
## probabilities: 0.362 0.367 0.170 0.088 0.012
## left son=6 (39298 obs) right son=7 (69518 obs)
## Primary splits:
## reimbursement2008 < 3065 to the left, improve=2010.3080, (0 missing)
## bucket2008 < 1.5 to the left, improve=1980.9770, (0 missing)
## kidney < 0.5 to the left, improve=1416.9220, (0 missing)
## diabetes < 0.5 to the left, improve=1236.1460, (0 missing)
## heart.failure < 0.5 to the left, improve= 976.9427, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.989, adj=0.969, (0 split)
## ihd < 0.5 to the left, agree=0.659, adj=0.056, (0 split)
## diabetes < 0.5 to the left, agree=0.641, adj=0.006, (0 split)
##
## Node number 6: 39298 observations
## predicted class=B1 expected loss=0.4797445 P(node) =0.1430043
## class counts: 20445 12134 4756 1782 181
## probabilities: 0.520 0.309 0.121 0.045 0.005
##
## Node number 7: 69518 observations, complexity param=0.01395884
## predicted class=B2 expected loss=0.5995138 P(node) =0.2529739
## class counts: 18972 27841 13728 7809 1168
## probabilities: 0.273 0.400 0.197 0.112 0.017
## left son=14 (15717 obs) right son=15 (53801 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=646.4740, (0 missing)
## kidney < 0.5 to the left, improve=604.0313, (0 missing)
## arthritis < 0.5 to the left, improve=501.1263, (0 missing)
## ihd < 0.5 to the left, improve=427.9009, (0 missing)
## heart.failure < 0.5 to the left, improve=380.0080, (0 missing)
##
## Node number 14: 15717 observations
## predicted class=B1 expected loss=0.5704651 P(node) =0.0571937
## class counts: 6751 5490 2365 999 112
## probabilities: 0.430 0.349 0.150 0.064 0.007
##
## Node number 15: 53801 observations
## predicted class=B2 expected loss=0.5845616 P(node) =0.1957802
## class counts: 12221 22351 11363 6810 1056
## probabilities: 0.227 0.415 0.211 0.127 0.020
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) reimbursement2008< 1565 165987 20938 B1 (0.87 0.074 0.037 0.014 0.0014) *
## 3) reimbursement2008>=1565 108816 68841 B2 (0.36 0.37 0.17 0.088 0.012)
## 6) reimbursement2008< 3065 39298 18853 B1 (0.52 0.31 0.12 0.045 0.0046) *
## 7) reimbursement2008>=3065 69518 41677 B2 (0.27 0.4 0.2 0.11 0.017)
## 14) diabetes< 0.5 15717 8966 B1 (0.43 0.35 0.15 0.064 0.0071) *
## 15) diabetes>=0.5 53801 31450 B2 (0.23 0.42 0.21 0.13 0.02) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 172245 12221 0 0 0
## B2 29908 22351 0 0 0
## B3 13223 11363 0 0 0
## B4 5096 6810 0 0 0
## B5 530 1056 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7081291 NA 0.7064253 0.7098285 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114902 8076 0 0 0
## B2 20130 14710 0 0 0
## B3 8749 7641 0 0 0
## B4 3409 4528 0 0 0
## B5 382 675 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.074814e-01 NA 7.053922e-01 7.095639e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 8.205962e-244 NaN
## model_id model_method
## 1 All.X.lser.no.cp.opt.rpart rpart
## feats
## 1 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 3 107.61 15.544
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.7086604 0.7064253 0.7098285
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 0.3140169 0.7074814 0.7053922
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.7095639 NA 0 0.001014054
## max.KappaSD.fit
## 1 0.004809558
## [1] "fitting model: All.X.lser.no.cp.4015.rpart"
## [1] " indep_vars: age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008"
## + Fold1: cp=5e-05
## - Fold1: cp=5e-05
## + Fold2: cp=5e-05
## - Fold2: cp=5e-05
## + Fold3: cp=5e-05
## - Fold3: cp=5e-05
## + Fold4: cp=5e-05
## - Fold4: cp=5e-05
## + Fold5: cp=5e-05
## - Fold5: cp=5e-05
## Aggregating results
## Fitting final model on full training set## Warning: labs do not fit even at cex 0.15, there may be some overplotting## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 4.908841e-02 0 1.0000000
## 2 1.395884e-02 2 0.9018232
## 3 4.350377e-03 3 0.8878643
## 4 3.597640e-03 4 0.8835140
## 5 9.879673e-04 5 0.8799163
## 6 7.859460e-04 10 0.8748685
## 7 6.789392e-04 11 0.8740826
## 8 4.649258e-04 14 0.8720458
## 9 4.372516e-04 15 0.8715809
## 10 2.988809e-04 19 0.8698319
## 11 2.767415e-04 20 0.8695330
## 12 2.435326e-04 21 0.8692562
## 13 2.036818e-04 22 0.8690127
## 14 1.881842e-04 28 0.8677729
## 15 1.826494e-04 29 0.8675847
## 16 1.605101e-04 31 0.8672194
## 17 1.439056e-04 33 0.8668984
## 18 1.411382e-04 37 0.8663228
## 19 1.217663e-04 42 0.8655922
## 20 1.162314e-04 45 0.8652269
## 21 1.129105e-04 47 0.8649944
## 22 1.051618e-04 52 0.8644299
## 23 9.962695e-05 57 0.8638764
## 24 9.409212e-05 68 0.8627473
## 25 8.855729e-05 74 0.8621827
## 26 8.302246e-05 83 0.8613746
## 27 7.748763e-05 85 0.8612086
## 28 7.379774e-05 97 0.8602455
## 29 7.195280e-05 101 0.8599134
## 30 6.918538e-05 114 0.8588729
## 31 6.641797e-05 122 0.8583194
## 32 6.272808e-05 154 0.8560612
## 33 6.167383e-05 158 0.8557955
## 34 6.088314e-05 166 0.8552752
## 35 5.811572e-05 179 0.8544340
## 36 5.534831e-05 183 0.8542015
## 37 5.313437e-05 228 0.8516001
## 38 5.258089e-05 233 0.8513344
## 39 5.165842e-05 237 0.8511241
## 40 5.000000e-05 254 0.8501832
##
## Variable importance
## reimbursement2008 bucket2008 diabetes ihd
## 31 20 13 13
## heart.failure kidney arthritis
## 11 9 1
##
## Node number 1: 274803 observations, complexity param=0.04908841
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (165987 obs) right son=3 (108816 obs)
## Primary splits:
## reimbursement2008 < 1565 to the left, improve=24395.14, (0 missing)
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## ihd < 0.5 to the left, improve=16291.74, (0 missing)
## diabetes < 0.5 to the left, improve=16041.26, (0 missing)
## heart.failure < 0.5 to the left, improve=12498.16, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.861, adj=0.650, (0 split)
## ihd < 0.5 to the left, agree=0.792, adj=0.474, (0 split)
## diabetes < 0.5 to the left, agree=0.785, adj=0.456, (0 split)
## heart.failure < 0.5 to the left, agree=0.762, adj=0.399, (0 split)
## kidney < 0.5 to the left, agree=0.731, adj=0.321, (0 split)
##
## Node number 2: 165987 observations
## predicted class=B1 expected loss=0.1261424 P(node) =0.6040218
## class counts: 145049 12284 6102 2315 237
## probabilities: 0.874 0.074 0.037 0.014 0.001
##
## Node number 3: 108816 observations, complexity param=0.04908841
## predicted class=B2 expected loss=0.6326367 P(node) =0.3959782
## class counts: 39417 39975 18484 9591 1349
## probabilities: 0.362 0.367 0.170 0.088 0.012
## left son=6 (39298 obs) right son=7 (69518 obs)
## Primary splits:
## reimbursement2008 < 3065 to the left, improve=2010.3080, (0 missing)
## bucket2008 < 1.5 to the left, improve=1980.9770, (0 missing)
## kidney < 0.5 to the left, improve=1416.9220, (0 missing)
## diabetes < 0.5 to the left, improve=1236.1460, (0 missing)
## heart.failure < 0.5 to the left, improve= 976.9427, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.989, adj=0.969, (0 split)
## ihd < 0.5 to the left, agree=0.659, adj=0.056, (0 split)
## diabetes < 0.5 to the left, agree=0.641, adj=0.006, (0 split)
##
## Node number 6: 39298 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.4797445 P(node) =0.1430043
## class counts: 20445 12134 4756 1782 181
## probabilities: 0.520 0.309 0.121 0.045 0.005
## left son=12 (20077 obs) right son=13 (19221 obs)
## Primary splits:
## reimbursement2008 < 2175 to the left, improve=192.7592, (0 missing)
## diabetes < 0.5 to the left, improve=155.3521, (0 missing)
## ihd < 0.5 to the left, improve=114.8541, (0 missing)
## arthritis < 0.5 to the left, improve=114.6837, (0 missing)
## kidney < 0.5 to the left, improve=108.9096, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.542, adj=0.063, (0 split)
## arthritis < 0.5 to the left, agree=0.539, adj=0.058, (0 split)
## ihd < 0.5 to the left, agree=0.534, adj=0.048, (0 split)
## kidney < 0.5 to the left, agree=0.532, adj=0.044, (0 split)
## diabetes < 0.5 to the left, agree=0.532, adj=0.043, (0 split)
##
## Node number 7: 69518 observations, complexity param=0.01395884
## predicted class=B2 expected loss=0.5995138 P(node) =0.2529739
## class counts: 18972 27841 13728 7809 1168
## probabilities: 0.273 0.400 0.197 0.112 0.017
## left son=14 (15717 obs) right son=15 (53801 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=646.4740, (0 missing)
## kidney < 0.5 to the left, improve=604.0313, (0 missing)
## arthritis < 0.5 to the left, improve=501.1263, (0 missing)
## ihd < 0.5 to the left, improve=427.9009, (0 missing)
## heart.failure < 0.5 to the left, improve=380.0080, (0 missing)
##
## Node number 12: 20077 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.4247148 P(node) =0.07305961
## class counts: 11550 5416 2200 834 77
## probabilities: 0.575 0.270 0.110 0.042 0.004
## left son=24 (8826 obs) right son=25 (11251 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=62.34344, (0 missing)
## kidney < 0.5 to the left, improve=42.15624, (0 missing)
## ihd < 0.5 to the left, improve=40.01287, (0 missing)
## heart.failure < 0.5 to the left, improve=36.00697, (0 missing)
## arthritis < 0.5 to the left, improve=33.77686, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.588, adj=0.062, (0 split)
##
## Node number 13: 19221 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.5372249 P(node) =0.06994465
## class counts: 8895 6718 2556 948 104
## probabilities: 0.463 0.350 0.133 0.049 0.005
## left son=26 (7137 obs) right son=27 (12084 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=71.31724, (0 missing)
## arthritis < 0.5 to the left, improve=61.00585, (0 missing)
## ihd < 0.5 to the left, improve=55.20411, (0 missing)
## heart.failure < 0.5 to the left, improve=52.20163, (0 missing)
## kidney < 0.5 to the left, improve=49.73230, (0 missing)
##
## Node number 14: 15717 observations, complexity param=0.004350377
## predicted class=B1 expected loss=0.5704651 P(node) =0.0571937
## class counts: 6751 5490 2365 999 112
## probabilities: 0.430 0.349 0.150 0.064 0.007
## left son=28 (13123 obs) right son=29 (2594 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=130.12270, (0 missing)
## arthritis < 0.5 to the left, improve=125.41530, (0 missing)
## ihd < 0.5 to the left, improve= 80.76118, (0 missing)
## depression < 0.5 to the left, improve= 61.32779, (0 missing)
## osteoporosis < 0.5 to the left, improve= 44.50253, (0 missing)
##
## Node number 15: 53801 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.5845616 P(node) =0.1957802
## class counts: 12221 22351 11363 6810 1056
## probabilities: 0.227 0.415 0.211 0.127 0.020
## left son=30 (25067 obs) right son=31 (28734 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=408.9756, (0 missing)
## reimbursement2008 < 15395 to the left, improve=327.1281, (0 missing)
## bucket2008 < 3.5 to the left, improve=313.8191, (0 missing)
## arthritis < 0.5 to the left, improve=266.5595, (0 missing)
## heart.failure < 0.5 to the left, improve=209.4718, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8365 to the left, agree=0.666, adj=0.282, (0 split)
## bucket2008 < 2.5 to the left, agree=0.664, adj=0.279, (0 split)
## heart.failure < 0.5 to the left, agree=0.628, adj=0.201, (0 split)
## copd < 0.5 to the left, agree=0.595, adj=0.132, (0 split)
## ihd < 0.5 to the left, agree=0.575, adj=0.089, (0 split)
##
## Node number 24: 8826 observations
## predicted class=B1 expected loss=0.3716293 P(node) =0.03211755
## class counts: 5546 2137 805 312 26
## probabilities: 0.628 0.242 0.091 0.035 0.003
##
## Node number 25: 11251 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.4663585 P(node) =0.04094206
## class counts: 6004 3279 1395 522 51
## probabilities: 0.534 0.291 0.124 0.046 0.005
## left son=50 (9007 obs) right son=51 (2244 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=18.86048, (0 missing)
## heart.failure < 0.5 to the left, improve=17.29926, (0 missing)
## arthritis < 0.5 to the left, improve=16.91283, (0 missing)
## reimbursement2008 < 1875 to the left, improve=16.48954, (0 missing)
## cancer < 0.5 to the left, improve=14.98495, (0 missing)
##
## Node number 26: 7137 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.470786 P(node) =0.02597133
## class counts: 3777 2233 794 300 33
## probabilities: 0.529 0.313 0.111 0.042 0.005
## left son=52 (5554 obs) right son=53 (1583 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=24.840370, (0 missing)
## depression < 0.5 to the left, improve=16.217060, (0 missing)
## ihd < 0.5 to the left, improve=13.895180, (0 missing)
## copd < 0.5 to the left, improve=12.688930, (0 missing)
## kidney < 0.5 to the left, improve= 9.728645, (0 missing)
##
## Node number 27: 12084 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.5764647 P(node) =0.04397332
## class counts: 5118 4485 1762 648 71
## probabilities: 0.424 0.371 0.146 0.054 0.006
## left son=54 (8413 obs) right son=55 (3671 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=27.83165, (0 missing)
## heart.failure < 0.5 to the left, improve=26.70933, (0 missing)
## ihd < 0.5 to the left, improve=24.37311, (0 missing)
## kidney < 0.5 to the left, improve=22.60183, (0 missing)
## reimbursement2008 < 2655 to the left, improve=21.75660, (0 missing)
##
## Node number 28: 13123 observations, complexity param=0.00359764
## predicted class=B1 expected loss=0.5360055 P(node) =0.04775421
## class counts: 6089 4435 1751 763 85
## probabilities: 0.464 0.338 0.133 0.058 0.006
## left son=56 (9625 obs) right son=57 (3498 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=126.28480, (0 missing)
## ihd < 0.5 to the left, improve= 70.76778, (0 missing)
## depression < 0.5 to the left, improve= 68.94332, (0 missing)
## osteoporosis < 0.5 to the left, improve= 46.31934, (0 missing)
## heart.failure < 0.5 to the left, improve= 30.26771, (0 missing)
##
## Node number 29: 2594 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.5932922 P(node) =0.009439489
## class counts: 662 1055 614 236 27
## probabilities: 0.255 0.407 0.237 0.091 0.010
## left son=58 (1000 obs) right son=59 (1594 obs)
## Primary splits:
## reimbursement2008 < 5770 to the left, improve=8.464458, (0 missing)
## arthritis < 0.5 to the left, improve=7.371565, (0 missing)
## ihd < 0.5 to the left, improve=5.410820, (0 missing)
## copd < 0.5 to the left, improve=5.301788, (0 missing)
## heart.failure < 0.5 to the left, improve=3.070575, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.823, adj=0.542, (0 split)
##
## Node number 30: 25067 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.57091 P(node) =0.09121807
## class counts: 7517 10756 4691 1917 186
## probabilities: 0.300 0.429 0.187 0.076 0.007
## left son=60 (15178 obs) right son=61 (9889 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=169.25970, (0 missing)
## cancer < 0.5 to the left, improve= 99.57556, (0 missing)
## ihd < 0.5 to the left, improve= 68.28883, (0 missing)
## depression < 0.5 to the left, improve= 61.94482, (0 missing)
## heart.failure < 0.5 to the left, improve= 42.19646, (0 missing)
## Surrogate splits:
## reimbursement2008 < 66495 to the left, agree=0.606, adj=0.001, (0 split)
##
## Node number 31: 28734 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.5964711 P(node) =0.1045622
## class counts: 4704 11595 6672 4893 870
## probabilities: 0.164 0.404 0.232 0.170 0.030
## left son=62 (16249 obs) right son=63 (12485 obs)
## Primary splits:
## reimbursement2008 < 15395 to the left, improve=177.49270, (0 missing)
## bucket2008 < 3.5 to the left, improve=170.28940, (0 missing)
## arthritis < 0.5 to the left, improve=101.31920, (0 missing)
## heart.failure < 0.5 to the left, improve= 62.82321, (0 missing)
## ihd < 0.5 to the left, improve= 55.35075, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.924, adj=0.826, (0 split)
## copd < 0.5 to the left, agree=0.609, adj=0.101, (0 split)
## stroke < 0.5 to the left, agree=0.605, adj=0.091, (0 split)
## cancer < 0.5 to the left, agree=0.580, adj=0.033, (0 split)
## alzheimers < 0.5 to the left, agree=0.569, adj=0.008, (0 split)
##
## Node number 50: 9007 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.4490951 P(node) =0.03277621
## class counts: 4962 2540 1087 378 40
## probabilities: 0.551 0.282 0.121 0.042 0.004
## left son=100 (4935 obs) right son=101 (4072 obs)
## Primary splits:
## reimbursement2008 < 1875 to the left, improve=14.670650, (0 missing)
## cancer < 0.5 to the left, improve=12.077140, (0 missing)
## arthritis < 0.5 to the left, improve= 9.470091, (0 missing)
## heart.failure < 0.5 to the left, improve= 7.308909, (0 missing)
## depression < 0.5 to the left, improve= 6.801973, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.549, adj=0.003, (0 split)
## age < 29.5 to the right, agree=0.548, adj=0.001, (0 split)
##
## Node number 51: 2244 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5356506 P(node) =0.00816585
## class counts: 1042 739 308 144 11
## probabilities: 0.464 0.329 0.137 0.064 0.005
## left son=102 (992 obs) right son=103 (1252 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=7.795458, (0 missing)
## arthritis < 0.5 to the left, improve=7.027320, (0 missing)
## ihd < 0.5 to the left, improve=4.964222, (0 missing)
## reimbursement2008 < 1735 to the left, improve=4.132280, (0 missing)
## cancer < 0.5 to the left, improve=3.835396, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.565, adj=0.016, (0 split)
## age < 33.5 to the left, agree=0.559, adj=0.002, (0 split)
##
## Node number 52: 5554 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4449046 P(node) =0.02021084
## class counts: 3083 1647 580 217 27
## probabilities: 0.555 0.297 0.104 0.039 0.005
## left son=104 (2348 obs) right son=105 (3206 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=13.118310, (0 missing)
## depression < 0.5 to the left, improve=12.689550, (0 missing)
## kidney < 0.5 to the left, improve= 9.684755, (0 missing)
## copd < 0.5 to the left, improve= 9.145592, (0 missing)
## heart.failure < 0.5 to the left, improve= 8.228139, (0 missing)
## Surrogate splits:
## age < 28.5 to the left, agree=0.579, adj=0.004, (0 split)
## reimbursement2008 < 2185 to the left, agree=0.578, adj=0.001, (0 split)
##
## Node number 53: 1583 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5615919 P(node) =0.00576049
## class counts: 694 586 214 83 6
## probabilities: 0.438 0.370 0.135 0.052 0.004
## left son=106 (1525 obs) right son=107 (58 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=5.133391, (0 missing)
## reimbursement2008 < 2725 to the left, improve=3.164238, (0 missing)
## cancer < 0.5 to the left, improve=2.451745, (0 missing)
## copd < 0.5 to the left, improve=2.436381, (0 missing)
## depression < 0.5 to the left, improve=1.979459, (0 missing)
##
## Node number 54: 8413 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.5530726 P(node) =0.03061466
## class counts: 3760 2943 1225 438 47
## probabilities: 0.447 0.350 0.146 0.052 0.006
## left son=108 (4375 obs) right son=109 (4038 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=25.12070, (0 missing)
## ihd < 0.5 to the left, improve=19.50225, (0 missing)
## kidney < 0.5 to the left, improve=18.23799, (0 missing)
## depression < 0.5 to the left, improve=14.07225, (0 missing)
## reimbursement2008 < 2615 to the left, improve=12.21338, (0 missing)
## Surrogate splits:
## kidney < 0.5 to the left, agree=0.569, adj=0.103, (0 split)
## copd < 0.5 to the left, agree=0.568, adj=0.100, (0 split)
## alzheimers < 0.5 to the left, agree=0.546, adj=0.054, (0 split)
## ihd < 0.5 to the left, agree=0.544, adj=0.050, (0 split)
## stroke < 0.5 to the left, agree=0.536, adj=0.034, (0 split)
##
## Node number 55: 3671 observations, complexity param=0.0002988809
## predicted class=B2 expected loss=0.579951 P(node) =0.01335866
## class counts: 1358 1542 537 210 24
## probabilities: 0.370 0.420 0.146 0.057 0.007
## left son=110 (2068 obs) right son=111 (1603 obs)
## Primary splits:
## reimbursement2008 < 2665 to the left, improve=10.442080, (0 missing)
## cancer < 0.5 to the left, improve= 4.234333, (0 missing)
## ihd < 0.5 to the left, improve= 4.129116, (0 missing)
## kidney < 0.5 to the left, improve= 3.679214, (0 missing)
## copd < 0.5 to the left, improve= 3.281268, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.638, adj=0.170, (0 split)
## cancer < 0.5 to the left, agree=0.566, adj=0.006, (0 split)
## age < 26.5 to the right, agree=0.564, adj=0.001, (0 split)
## stroke < 0.5 to the left, agree=0.564, adj=0.001, (0 split)
##
## Node number 56: 9625 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.4874805 P(node) =0.03502509
## class counts: 4933 2954 1162 520 56
## probabilities: 0.513 0.307 0.121 0.054 0.006
## left son=112 (3135 obs) right son=113 (6490 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=51.18602, (0 missing)
## depression < 0.5 to the left, improve=46.82343, (0 missing)
## heart.failure < 0.5 to the left, improve=27.25528, (0 missing)
## osteoporosis < 0.5 to the left, improve=25.54800, (0 missing)
## reimbursement2008 < 6615 to the left, improve=12.84564, (0 missing)
##
## Node number 57: 3498 observations, complexity param=0.0004372516
## predicted class=B2 expected loss=0.5766152 P(node) =0.01272912
## class counts: 1156 1481 589 243 29
## probabilities: 0.330 0.423 0.168 0.069 0.008
## left son=114 (2340 obs) right son=115 (1158 obs)
## Primary splits:
## reimbursement2008 < 8525 to the left, improve=12.263650, (0 missing)
## depression < 0.5 to the left, improve=10.454350, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.052395, (0 missing)
## copd < 0.5 to the left, improve= 8.848663, (0 missing)
## ihd < 0.5 to the left, improve= 8.087092, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.979, adj=0.935, (0 split)
## kidney < 0.5 to the left, agree=0.692, adj=0.069, (0 split)
## stroke < 0.5 to the left, agree=0.680, adj=0.033, (0 split)
##
## Node number 58: 1000 observations
## predicted class=B2 expected loss=0.562 P(node) =0.00363897
## class counts: 296 438 191 70 5
## probabilities: 0.296 0.438 0.191 0.070 0.005
##
## Node number 59: 1594 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6129235 P(node) =0.005800519
## class counts: 366 617 423 166 22
## probabilities: 0.230 0.387 0.265 0.104 0.014
## left son=118 (1054 obs) right son=119 (540 obs)
## Primary splits:
## reimbursement2008 < 8645 to the right, improve=7.014383, (0 missing)
## arthritis < 0.5 to the left, improve=5.636989, (0 missing)
## bucket2008 < 2.5 to the right, improve=4.256675, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.245615, (0 missing)
## ihd < 0.5 to the left, improve=2.672736, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.949, adj=0.848, (0 split)
## age < 27.5 to the right, agree=0.662, adj=0.002, (0 split)
##
## Node number 60: 15178 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.6047569 P(node) =0.05523229
## class counts: 5388 5999 2649 1047 95
## probabilities: 0.355 0.395 0.175 0.069 0.006
## left son=120 (12572 obs) right son=121 (2606 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=92.65854, (0 missing)
## ihd < 0.5 to the left, improve=43.72992, (0 missing)
## depression < 0.5 to the left, improve=36.05906, (0 missing)
## heart.failure < 0.5 to the left, improve=30.26654, (0 missing)
## copd < 0.5 to the left, improve=25.73984, (0 missing)
##
## Node number 61: 9889 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5189605 P(node) =0.03598578
## class counts: 2129 4757 2042 870 91
## probabilities: 0.215 0.481 0.206 0.088 0.009
## left son=122 (5134 obs) right son=123 (4755 obs)
## Primary splits:
## depression < 0.5 to the left, improve=18.84327, (0 missing)
## cancer < 0.5 to the left, improve=17.45891, (0 missing)
## ihd < 0.5 to the left, improve=13.35120, (0 missing)
## reimbursement2008 < 9795 to the left, improve=12.33086, (0 missing)
## copd < 0.5 to the left, improve=12.26415, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.564, adj=0.093, (0 split)
## copd < 0.5 to the left, agree=0.546, adj=0.056, (0 split)
## reimbursement2008 < 5815 to the left, agree=0.542, adj=0.048, (0 split)
## age < 64.5 to the right, agree=0.537, adj=0.037, (0 split)
## bucket2008 < 2.5 to the left, agree=0.536, adj=0.036, (0 split)
##
## Node number 62: 16249 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.5619423 P(node) =0.05912963
## class counts: 3113 7118 3819 1946 253
## probabilities: 0.192 0.438 0.235 0.120 0.016
## left son=124 (9424 obs) right son=125 (6825 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=70.60653, (0 missing)
## cancer < 0.5 to the left, improve=30.24922, (0 missing)
## ihd < 0.5 to the left, improve=29.86941, (0 missing)
## reimbursement2008 < 5665 to the left, improve=23.89268, (0 missing)
## bucket2008 < 2.5 to the left, improve=21.55872, (0 missing)
##
## Node number 63: 12485 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.6414097 P(node) =0.04543255
## class counts: 1591 4477 2853 2947 617
## probabilities: 0.127 0.359 0.229 0.236 0.049
## left son=126 (5402 obs) right son=127 (7083 obs)
## Primary splits:
## arthritis < 0.5 to the right, improve=35.40534, (0 missing)
## cancer < 0.5 to the left, improve=26.78171, (0 missing)
## reimbursement2008 < 26625 to the left, improve=24.60405, (0 missing)
## depression < 0.5 to the left, improve=23.29796, (0 missing)
## heart.failure < 0.5 to the left, improve=17.01274, (0 missing)
## Surrogate splits:
## age < 28.5 to the left, agree=0.568, adj=0.002, (0 split)
## reimbursement2008 < 15435 to the left, agree=0.568, adj=0.001, (0 split)
##
## Node number 100: 4935 observations
## predicted class=B1 expected loss=0.4196555 P(node) =0.01795832
## class counts: 2864 1294 550 205 22
## probabilities: 0.580 0.262 0.111 0.042 0.004
##
## Node number 101: 4072 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.4847741 P(node) =0.01481789
## class counts: 2098 1246 537 173 18
## probabilities: 0.515 0.306 0.132 0.042 0.004
## left son=202 (3786 obs) right son=203 (286 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=5.937439, (0 missing)
## arthritis < 0.5 to the left, improve=5.625805, (0 missing)
## copd < 0.5 to the left, improve=3.348444, (0 missing)
## ihd < 0.5 to the left, improve=3.030239, (0 missing)
## heart.failure < 0.5 to the left, improve=2.851779, (0 missing)
##
## Node number 102: 992 observations
## predicted class=B1 expected loss=0.4808468 P(node) =0.003609859
## class counts: 515 292 126 57 2
## probabilities: 0.519 0.294 0.127 0.057 0.002
##
## Node number 103: 1252 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5790735 P(node) =0.004555991
## class counts: 527 447 182 87 9
## probabilities: 0.421 0.357 0.145 0.069 0.007
## left son=206 (904 obs) right son=207 (348 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=7.739842, (0 missing)
## age < 93.5 to the left, improve=3.754099, (0 missing)
## cancer < 0.5 to the left, improve=3.514161, (0 missing)
## reimbursement2008 < 1955 to the left, improve=3.377454, (0 missing)
## ihd < 0.5 to the left, improve=1.751139, (0 missing)
## Surrogate splits:
## age < 30.5 to the right, agree=0.724, adj=0.006, (0 split)
##
## Node number 104: 2348 observations
## predicted class=B1 expected loss=0.3973595 P(node) =0.008544303
## class counts: 1415 632 217 72 12
## probabilities: 0.603 0.269 0.092 0.031 0.005
##
## Node number 105: 3206 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4797255 P(node) =0.01166654
## class counts: 1668 1015 363 145 15
## probabilities: 0.520 0.317 0.113 0.045 0.005
## left son=210 (2325 obs) right son=211 (881 obs)
## Primary splits:
## depression < 0.5 to the left, improve=8.135493, (0 missing)
## kidney < 0.5 to the left, improve=5.219511, (0 missing)
## reimbursement2008 < 2785 to the left, improve=4.205524, (0 missing)
## heart.failure < 0.5 to the left, improve=3.201394, (0 missing)
## copd < 0.5 to the left, improve=3.002159, (0 missing)
## Surrogate splits:
## age < 29.5 to the right, agree=0.726, adj=0.003, (0 split)
##
## Node number 106: 1525 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5534426 P(node) =0.00554943
## class counts: 681 554 202 82 6
## probabilities: 0.447 0.363 0.132 0.054 0.004
## left son=212 (1438 obs) right son=213 (87 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.548424, (0 missing)
## reimbursement2008 < 2715 to the right, improve=2.513748, (0 missing)
## copd < 0.5 to the left, improve=1.973703, (0 missing)
## depression < 0.5 to the left, improve=1.853940, (0 missing)
## kidney < 0.5 to the left, improve=1.632947, (0 missing)
##
## Node number 107: 58 observations
## predicted class=B2 expected loss=0.4482759 P(node) =0.0002110603
## class counts: 13 32 12 1 0
## probabilities: 0.224 0.552 0.207 0.017 0.000
##
## Node number 108: 4375 observations, complexity param=0.0002435326
## predicted class=B1 expected loss=0.5074286 P(node) =0.0159205
## class counts: 2155 1478 555 170 17
## probabilities: 0.493 0.338 0.127 0.039 0.004
## left son=216 (3992 obs) right son=217 (383 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=10.015540, (0 missing)
## ihd < 0.5 to the left, improve= 9.488719, (0 missing)
## depression < 0.5 to the left, improve= 7.316301, (0 missing)
## reimbursement2008 < 2615 to the left, improve= 5.949976, (0 missing)
## copd < 0.5 to the left, improve= 5.117423, (0 missing)
##
## Node number 109: 4038 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.602526 P(node) =0.01469416
## class counts: 1605 1465 670 268 30
## probabilities: 0.397 0.363 0.166 0.066 0.007
## left son=218 (2819 obs) right son=219 (1219 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=10.392200, (0 missing)
## reimbursement2008 < 2455 to the left, improve= 6.028802, (0 missing)
## ihd < 0.5 to the left, improve= 5.795095, (0 missing)
## depression < 0.5 to the left, improve= 5.214940, (0 missing)
## stroke < 0.5 to the left, improve= 3.343262, (0 missing)
##
## Node number 110: 2068 observations, complexity param=0.0002767415
## predicted class=B1 expected loss=0.5918762 P(node) =0.007525391
## class counts: 844 817 280 117 10
## probabilities: 0.408 0.395 0.135 0.057 0.005
## left son=220 (517 obs) right son=221 (1551 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.581883, (0 missing)
## reimbursement2008 < 2305 to the left, improve=3.255344, (0 missing)
## cancer < 0.5 to the left, improve=3.097089, (0 missing)
## age < 54.5 to the left, improve=1.964830, (0 missing)
## kidney < 0.5 to the left, improve=1.730688, (0 missing)
##
## Node number 111: 1603 observations
## predicted class=B2 expected loss=0.547723 P(node) =0.00583327
## class counts: 514 725 257 93 14
## probabilities: 0.321 0.452 0.160 0.058 0.009
##
## Node number 112: 3135 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.3974482 P(node) =0.01140817
## class counts: 1889 825 298 113 10
## probabilities: 0.603 0.263 0.095 0.036 0.003
## left son=224 (2292 obs) right son=225 (843 obs)
## Primary splits:
## depression < 0.5 to the left, improve=19.892930, (0 missing)
## reimbursement2008 < 9505 to the right, improve=15.211730, (0 missing)
## bucket2008 < 2.5 to the right, improve=13.054300, (0 missing)
## osteoporosis < 0.5 to the left, improve=10.317040, (0 missing)
## age < 92.5 to the left, improve= 3.244996, (0 missing)
## Surrogate splits:
## reimbursement2008 < 60755 to the left, agree=0.731, adj=0.001, (0 split)
##
## Node number 113: 6490 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.5309707 P(node) =0.02361692
## class counts: 3044 2129 864 407 46
## probabilities: 0.469 0.328 0.133 0.063 0.007
## left son=226 (4266 obs) right son=227 (2224 obs)
## Primary splits:
## depression < 0.5 to the left, improve=22.130520, (0 missing)
## heart.failure < 0.5 to the left, improve=12.472230, (0 missing)
## osteoporosis < 0.5 to the left, improve=12.135520, (0 missing)
## reimbursement2008 < 6615 to the left, improve=10.028930, (0 missing)
## bucket2008 < 2.5 to the left, improve= 8.000565, (0 missing)
## Surrogate splits:
## age < 34.5 to the right, agree=0.658, adj=0.003, (0 split)
## reimbursement2008 < 115145 to the left, agree=0.658, adj=0.001, (0 split)
##
## Node number 114: 2340 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.542735 P(node) =0.008515191
## class counts: 720 1070 391 144 15
## probabilities: 0.308 0.457 0.167 0.062 0.006
## left son=228 (1359 obs) right son=229 (981 obs)
## Primary splits:
## reimbursement2008 < 4645 to the left, improve=5.782135, (0 missing)
## ihd < 0.5 to the left, improve=5.431632, (0 missing)
## depression < 0.5 to the left, improve=4.505952, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.336155, (0 missing)
## alzheimers < 0.5 to the left, improve=3.247654, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.613, adj=0.076, (0 split)
## copd < 0.5 to the left, agree=0.606, adj=0.059, (0 split)
## kidney < 0.5 to the left, agree=0.586, adj=0.013, (0 split)
## age < 91.5 to the left, agree=0.585, adj=0.011, (0 split)
## stroke < 0.5 to the left, agree=0.585, adj=0.009, (0 split)
##
## Node number 115: 1158 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.6234888 P(node) =0.004213928
## class counts: 436 411 198 99 14
## probabilities: 0.377 0.355 0.171 0.085 0.012
## left son=230 (714 obs) right son=231 (444 obs)
## Primary splits:
## copd < 0.5 to the left, improve=13.168040, (0 missing)
## depression < 0.5 to the left, improve= 8.948306, (0 missing)
## kidney < 0.5 to the left, improve= 6.276303, (0 missing)
## ihd < 0.5 to the left, improve= 5.293866, (0 missing)
## reimbursement2008 < 14980 to the left, improve= 4.056180, (0 missing)
## Surrogate splits:
## age < 94.5 to the left, agree=0.626, adj=0.025, (0 split)
## reimbursement2008 < 72745 to the left, agree=0.620, adj=0.009, (0 split)
##
## Node number 118: 1054 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6223909 P(node) =0.003835475
## class counts: 281 398 250 109 16
## probabilities: 0.267 0.378 0.237 0.103 0.015
## left son=236 (745 obs) right son=237 (309 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=8.492364, (0 missing)
## ihd < 0.5 to the left, improve=3.739184, (0 missing)
## depression < 0.5 to the left, improve=2.714506, (0 missing)
## copd < 0.5 to the left, improve=2.704564, (0 missing)
## reimbursement2008 < 67610 to the left, improve=2.665770, (0 missing)
## Surrogate splits:
## age < 29.5 to the right, agree=0.708, adj=0.003, (0 split)
##
## Node number 119: 540 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.5944444 P(node) =0.001965044
## class counts: 85 219 173 57 6
## probabilities: 0.157 0.406 0.320 0.106 0.011
## left son=238 (243 obs) right son=239 (297 obs)
## Primary splits:
## heart.failure < 0.5 to the right, improve=3.144781, (0 missing)
## reimbursement2008 < 7455 to the left, improve=1.665302, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.352183, (0 missing)
## age < 86.5 to the right, improve=1.232072, (0 missing)
## arthritis < 0.5 to the right, improve=1.028824, (0 missing)
## Surrogate splits:
## copd < 0.5 to the right, agree=0.604, adj=0.119, (0 split)
## kidney < 0.5 to the right, agree=0.585, adj=0.078, (0 split)
## stroke < 0.5 to the right, agree=0.583, adj=0.074, (0 split)
## arthritis < 0.5 to the right, agree=0.576, adj=0.058, (0 split)
## depression < 0.5 to the right, agree=0.572, adj=0.049, (0 split)
##
## Node number 120: 12572 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.613188 P(node) =0.04574914
## class counts: 4844 4863 2000 791 74
## probabilities: 0.385 0.387 0.159 0.063 0.006
## left son=240 (2617 obs) right son=241 (9955 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=36.80981, (0 missing)
## depression < 0.5 to the left, improve=36.47326, (0 missing)
## heart.failure < 0.5 to the left, improve=27.52215, (0 missing)
## copd < 0.5 to the left, improve=21.85222, (0 missing)
## reimbursement2008 < 8955 to the left, improve=19.34797, (0 missing)
##
## Node number 121: 2606 observations
## predicted class=B2 expected loss=0.5640829 P(node) =0.009483157
## class counts: 544 1136 649 256 21
## probabilities: 0.209 0.436 0.249 0.098 0.008
##
## Node number 122: 5134 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5190884 P(node) =0.01868247
## class counts: 1277 2469 936 412 40
## probabilities: 0.249 0.481 0.182 0.080 0.008
## left son=244 (4305 obs) right son=245 (829 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=12.348810, (0 missing)
## reimbursement2008 < 9985 to the left, improve=11.590150, (0 missing)
## bucket2008 < 2.5 to the left, improve= 7.979608, (0 missing)
## ihd < 0.5 to the left, improve= 7.512372, (0 missing)
## copd < 0.5 to the left, improve= 7.186891, (0 missing)
##
## Node number 123: 4755 observations
## predicted class=B2 expected loss=0.5188223 P(node) =0.0173033
## class counts: 852 2288 1106 458 51
## probabilities: 0.179 0.481 0.233 0.096 0.011
##
## Node number 124: 9424 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.5992148 P(node) =0.03429366
## class counts: 2192 3777 2139 1156 160
## probabilities: 0.233 0.401 0.227 0.123 0.017
## left son=248 (7786 obs) right son=249 (1638 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=31.06099, (0 missing)
## ihd < 0.5 to the left, improve=19.93184, (0 missing)
## depression < 0.5 to the left, improve=16.57581, (0 missing)
## reimbursement2008 < 6325 to the left, improve=12.91187, (0 missing)
## bucket2008 < 2.5 to the left, improve=10.82187, (0 missing)
##
## Node number 125: 6825 observations
## predicted class=B2 expected loss=0.5104762 P(node) =0.02483597
## class counts: 921 3341 1680 790 93
## probabilities: 0.135 0.490 0.246 0.116 0.014
##
## Node number 126: 5402 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5960755 P(node) =0.01965772
## class counts: 509 2182 1310 1186 215
## probabilities: 0.094 0.404 0.243 0.220 0.040
## left son=252 (3345 obs) right son=253 (2057 obs)
## Primary splits:
## reimbursement2008 < 34925 to the left, improve=14.212070, (0 missing)
## copd < 0.5 to the left, improve=10.384850, (0 missing)
## depression < 0.5 to the left, improve= 8.104595, (0 missing)
## cancer < 0.5 to the right, improve= 6.743072, (0 missing)
## heart.failure < 0.5 to the left, improve= 6.417519, (0 missing)
## Surrogate splits:
## bucket2008 < 4.5 to the left, agree=0.776, adj=0.413, (0 split)
##
## Node number 127: 7083 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.6759848 P(node) =0.02577483
## class counts: 1082 2295 1543 1761 402
## probabilities: 0.153 0.324 0.218 0.249 0.057
## left son=254 (5298 obs) right son=255 (1785 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=21.09129, (0 missing)
## depression < 0.5 to the left, improve=19.29947, (0 missing)
## reimbursement2008 < 26625 to the left, improve=15.18952, (0 missing)
## copd < 0.5 to the left, improve=14.68870, (0 missing)
## heart.failure < 0.5 to the left, improve=12.81802, (0 missing)
##
## Node number 202: 3786 observations
## predicted class=B1 expected loss=0.4772847 P(node) =0.01377714
## class counts: 1979 1131 501 157 18
## probabilities: 0.523 0.299 0.132 0.041 0.005
##
## Node number 203: 286 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5839161 P(node) =0.001040746
## class counts: 119 115 36 16 0
## probabilities: 0.416 0.402 0.126 0.056 0.000
## left son=406 (128 obs) right son=407 (158 obs)
## Primary splits:
## age < 73.5 to the left, improve=2.9724540, (0 missing)
## reimbursement2008 < 2005 to the left, improve=1.9802050, (0 missing)
## depression < 0.5 to the left, improve=0.5460014, (0 missing)
## alzheimers < 0.5 to the right, improve=0.4144954, (0 missing)
## ihd < 0.5 to the left, improve=0.3767582, (0 missing)
## Surrogate splits:
## reimbursement2008 < 1945 to the left, agree=0.580, adj=0.063, (0 split)
## arthritis < 0.5 to the right, agree=0.563, adj=0.023, (0 split)
##
## Node number 206: 904 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5376106 P(node) =0.003289629
## class counts: 418 304 119 57 6
## probabilities: 0.462 0.336 0.132 0.063 0.007
## left son=412 (270 obs) right son=413 (634 obs)
## Primary splits:
## reimbursement2008 < 1735 to the left, improve=3.8438620, (0 missing)
## age < 93.5 to the left, improve=3.6681650, (0 missing)
## ihd < 0.5 to the left, improve=3.2669730, (0 missing)
## cancer < 0.5 to the left, improve=3.0869480, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7446912, (0 missing)
## Surrogate splits:
## age < 29 to the left, agree=0.702, adj=0.004, (0 split)
##
## Node number 207: 348 observations
## predicted class=B2 expected loss=0.5890805 P(node) =0.001266362
## class counts: 109 143 63 30 3
## probabilities: 0.313 0.411 0.181 0.086 0.009
##
## Node number 210: 2325 observations
## predicted class=B1 expected loss=0.4541935 P(node) =0.008460606
## class counts: 1269 700 245 99 12
## probabilities: 0.546 0.301 0.105 0.043 0.005
##
## Node number 211: 881 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5471056 P(node) =0.003205933
## class counts: 399 315 118 46 3
## probabilities: 0.453 0.358 0.134 0.052 0.003
## left son=422 (763 obs) right son=423 (118 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=3.122415, (0 missing)
## age < 78.5 to the right, improve=2.656467, (0 missing)
## reimbursement2008 < 2205 to the right, improve=1.600090, (0 missing)
## stroke < 0.5 to the left, improve=1.074836, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.071176, (0 missing)
##
## Node number 212: 1438 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5452017 P(node) =0.00523284
## class counts: 654 515 187 76 6
## probabilities: 0.455 0.358 0.130 0.053 0.004
## left son=424 (495 obs) right son=425 (943 obs)
## Primary splits:
## reimbursement2008 < 2715 to the right, improve=2.835023, (0 missing)
## kidney < 0.5 to the left, improve=1.879898, (0 missing)
## copd < 0.5 to the left, improve=1.857999, (0 missing)
## age < 40.5 to the right, improve=1.802592, (0 missing)
## heart.failure < 0.5 to the left, improve=1.761837, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=0.713, adj=0.166, (0 split)
## age < 28.5 to the left, agree=0.656, adj=0.002, (0 split)
##
## Node number 213: 87 observations
## predicted class=B2 expected loss=0.5517241 P(node) =0.0003165904
## class counts: 27 39 15 6 0
## probabilities: 0.310 0.448 0.172 0.069 0.000
##
## Node number 216: 3992 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.495491 P(node) =0.01452677
## class counts: 2014 1315 497 153 13
## probabilities: 0.505 0.329 0.124 0.038 0.003
## left son=432 (1265 obs) right son=433 (2727 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=7.867939, (0 missing)
## depression < 0.5 to the left, improve=6.016589, (0 missing)
## copd < 0.5 to the left, improve=5.402587, (0 missing)
## kidney < 0.5 to the left, improve=3.916699, (0 missing)
## reimbursement2008 < 2615 to the left, improve=3.836002, (0 missing)
##
## Node number 217: 383 observations, complexity param=0.0001826494
## predicted class=B2 expected loss=0.5744125 P(node) =0.001393726
## class counts: 141 163 58 17 4
## probabilities: 0.368 0.426 0.151 0.044 0.010
## left son=434 (238 obs) right son=435 (145 obs)
## Primary splits:
## reimbursement2008 < 2705 to the left, improve=4.9624930, (0 missing)
## depression < 0.5 to the left, improve=3.2303380, (0 missing)
## age < 67.5 to the right, improve=2.3511250, (0 missing)
## ihd < 0.5 to the left, improve=1.5735720, (0 missing)
## bucket2008 < 1.5 to the left, improve=0.9813303, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.681, adj=0.159, (0 split)
## age < 45.5 to the right, agree=0.624, adj=0.007, (0 split)
##
## Node number 218: 2819 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.5746719 P(node) =0.01025826
## class counts: 1199 980 439 183 18
## probabilities: 0.425 0.348 0.156 0.065 0.006
## left son=436 (635 obs) right son=437 (2184 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=6.072389, (0 missing)
## reimbursement2008 < 2325 to the left, improve=3.797765, (0 missing)
## age < 40.5 to the right, improve=3.110525, (0 missing)
## depression < 0.5 to the left, improve=2.993563, (0 missing)
## stroke < 0.5 to the left, improve=2.412511, (0 missing)
##
## Node number 219: 1219 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.6021329 P(node) =0.004435905
## class counts: 406 485 231 85 12
## probabilities: 0.333 0.398 0.189 0.070 0.010
## left son=438 (613 obs) right son=439 (606 obs)
## Primary splits:
## reimbursement2008 < 2615 to the left, improve=4.2080810, (0 missing)
## age < 98.5 to the right, improve=2.1482090, (0 missing)
## depression < 0.5 to the left, improve=1.6601240, (0 missing)
## stroke < 0.5 to the left, improve=0.8099205, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.7434054, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.579, adj=0.153, (0 split)
## depression < 0.5 to the left, agree=0.523, adj=0.041, (0 split)
## stroke < 0.5 to the left, agree=0.522, adj=0.038, (0 split)
## age < 65.5 to the right, agree=0.519, adj=0.033, (0 split)
## cancer < 0.5 to the left, agree=0.514, adj=0.021, (0 split)
##
## Node number 220: 517 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5299807 P(node) =0.001881348
## class counts: 243 191 57 25 1
## probabilities: 0.470 0.369 0.110 0.048 0.002
## left son=440 (143 obs) right son=441 (374 obs)
## Primary splits:
## reimbursement2008 < 2295 to the left, improve=6.0966680, (0 missing)
## cancer < 0.5 to the left, improve=2.5628030, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.9493160, (0 missing)
## age < 44.5 to the right, improve=1.5968610, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9005685, (0 missing)
## Surrogate splits:
## age < 98.5 to the right, agree=0.729, adj=0.021, (0 split)
##
## Node number 221: 1551 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5963894 P(node) =0.005644043
## class counts: 601 626 223 92 9
## probabilities: 0.387 0.404 0.144 0.059 0.006
## left son=442 (18 obs) right son=443 (1533 obs)
## Primary splits:
## age < 35 to the left, improve=3.0170030, (0 missing)
## kidney < 0.5 to the left, improve=2.3281310, (0 missing)
## cancer < 0.5 to the left, improve=1.5502140, (0 missing)
## stroke < 0.5 to the left, improve=1.1903410, (0 missing)
## copd < 0.5 to the left, improve=0.9727402, (0 missing)
##
## Node number 224: 2292 observations
## predicted class=B1 expected loss=0.3582024 P(node) =0.00834052
## class counts: 1471 549 183 79 10
## probabilities: 0.642 0.240 0.080 0.034 0.004
##
## Node number 225: 843 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5041518 P(node) =0.003067652
## class counts: 418 276 115 34 0
## probabilities: 0.496 0.327 0.136 0.040 0.000
## left son=450 (810 obs) right son=451 (33 obs)
## Primary splits:
## age < 92.5 to the left, improve=5.7055350, (0 missing)
## reimbursement2008 < 11540 to the right, improve=5.6370950, (0 missing)
## bucket2008 < 2.5 to the right, improve=2.9317810, (0 missing)
## stroke < 0.5 to the left, improve=0.7284423, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3506867, (0 missing)
##
## Node number 226: 4266 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.4946085 P(node) =0.01552385
## class counts: 2156 1343 503 238 26
## probabilities: 0.505 0.315 0.118 0.056 0.006
## left son=452 (3304 obs) right son=453 (962 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=10.212680, (0 missing)
## reimbursement2008 < 5905 to the right, improve= 9.673580, (0 missing)
## bucket2008 < 2.5 to the right, improve= 7.844764, (0 missing)
## heart.failure < 0.5 to the left, improve= 6.371374, (0 missing)
## age < 62.5 to the left, improve= 3.683231, (0 missing)
##
## Node number 227: 2224 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.6007194 P(node) =0.00809307
## class counts: 888 786 361 169 20
## probabilities: 0.399 0.353 0.162 0.076 0.009
## left son=454 (1518 obs) right son=455 (706 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=6.746609, (0 missing)
## heart.failure < 0.5 to the left, improve=4.569316, (0 missing)
## reimbursement2008 < 10710 to the left, improve=3.711923, (0 missing)
## age < 39.5 to the right, improve=3.285727, (0 missing)
## bucket2008 < 2.5 to the left, improve=2.661027, (0 missing)
## Surrogate splits:
## reimbursement2008 < 14380 to the left, agree=0.714, adj=0.101, (0 split)
## bucket2008 < 3.5 to the left, agree=0.708, adj=0.081, (0 split)
## age < 98.5 to the left, agree=0.684, adj=0.004, (0 split)
##
## Node number 228: 1359 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5548197 P(node) =0.004945361
## class counts: 467 605 203 76 8
## probabilities: 0.344 0.445 0.149 0.056 0.006
## left son=456 (440 obs) right son=457 (919 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.300387, (0 missing)
## alzheimers < 0.5 to the left, improve=2.001264, (0 missing)
## reimbursement2008 < 3265 to the left, improve=1.998939, (0 missing)
## depression < 0.5 to the left, improve=1.755319, (0 missing)
## heart.failure < 0.5 to the left, improve=1.574681, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3095 to the left, agree=0.678, adj=0.007, (0 split)
## age < 29.5 to the left, agree=0.678, adj=0.005, (0 split)
##
## Node number 229: 981 observations
## predicted class=B2 expected loss=0.5259939 P(node) =0.00356983
## class counts: 253 465 188 68 7
## probabilities: 0.258 0.474 0.192 0.069 0.007
##
## Node number 230: 714 observations, complexity param=0.0001881842
## predicted class=B1 expected loss=0.5546218 P(node) =0.002598225
## class counts: 318 239 91 61 5
## probabilities: 0.445 0.335 0.127 0.085 0.007
## left son=460 (412 obs) right son=461 (302 obs)
## Primary splits:
## depression < 0.5 to the left, improve=8.699660, (0 missing)
## age < 92.5 to the right, improve=3.253447, (0 missing)
## reimbursement2008 < 14980 to the left, improve=2.826720, (0 missing)
## bucket2008 < 3.5 to the left, improve=2.191697, (0 missing)
## kidney < 0.5 to the left, improve=2.037790, (0 missing)
## Surrogate splits:
## reimbursement2008 < 32685 to the left, agree=0.591, adj=0.033, (0 split)
## age < 35.5 to the right, agree=0.583, adj=0.013, (0 split)
##
## Node number 231: 444 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.6126126 P(node) =0.001615703
## class counts: 118 172 107 38 9
## probabilities: 0.266 0.387 0.241 0.086 0.020
## left son=462 (282 obs) right son=463 (162 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=3.735228, (0 missing)
## kidney < 0.5 to the left, improve=3.274615, (0 missing)
## reimbursement2008 < 68975 to the right, improve=3.185223, (0 missing)
## ihd < 0.5 to the left, improve=3.085645, (0 missing)
## age < 76.5 to the left, improve=1.652811, (0 missing)
## Surrogate splits:
## age < 95.5 to the left, agree=0.644, adj=0.025, (0 split)
## reimbursement2008 < 8635 to the right, agree=0.637, adj=0.006, (0 split)
##
## Node number 236: 745 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6228188 P(node) =0.002711033
## class counts: 232 281 150 72 10
## probabilities: 0.311 0.377 0.201 0.097 0.013
## left son=472 (159 obs) right son=473 (586 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.002920, (0 missing)
## reimbursement2008 < 58135 to the left, improve=2.259882, (0 missing)
## depression < 0.5 to the left, improve=2.111862, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.991400, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.920660, (0 missing)
##
## Node number 237: 309 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6213592 P(node) =0.001124442
## class counts: 49 117 100 37 6
## probabilities: 0.159 0.379 0.324 0.120 0.019
## left son=474 (237 obs) right son=475 (72 obs)
## Primary splits:
## reimbursement2008 < 10960 to the right, improve=2.966323, (0 missing)
## alzheimers < 0.5 to the right, improve=1.571780, (0 missing)
## age < 90.5 to the left, improve=1.407411, (0 missing)
## copd < 0.5 to the left, improve=1.306020, (0 missing)
## stroke < 0.5 to the left, improve=0.907593, (0 missing)
##
## Node number 238: 243 observations
## predicted class=B2 expected loss=0.526749 P(node) =0.0008842698
## class counts: 33 115 67 24 4
## probabilities: 0.136 0.473 0.276 0.099 0.016
##
## Node number 239: 297 observations, complexity param=6.167383e-05
## predicted class=B3 expected loss=0.6430976 P(node) =0.001080774
## class counts: 52 104 106 33 2
## probabilities: 0.175 0.350 0.357 0.111 0.007
## left son=478 (226 obs) right son=479 (71 obs)
## Primary splits:
## depression < 0.5 to the left, improve=1.480103, (0 missing)
## reimbursement2008 < 6875 to the right, improve=1.383473, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.357727, (0 missing)
## age < 54 to the left, improve=1.263809, (0 missing)
## alzheimers < 0.5 to the right, improve=1.096200, (0 missing)
##
## Node number 240: 2617 observations, complexity param=0.0001439056
## predicted class=B1 expected loss=0.5257929 P(node) =0.009523186
## class counts: 1241 884 351 127 14
## probabilities: 0.474 0.338 0.134 0.049 0.005
## left son=480 (403 obs) right son=481 (2214 obs)
## Primary splits:
## reimbursement2008 < 9400 to the right, improve=12.428110, (0 missing)
## bucket2008 < 2.5 to the right, improve= 8.843694, (0 missing)
## depression < 0.5 to the left, improve= 8.588030, (0 missing)
## osteoporosis < 0.5 to the left, improve= 8.405901, (0 missing)
## alzheimers < 0.5 to the left, improve= 4.036896, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.947, adj=0.658, (0 split)
##
## Node number 241: 9955 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.6003014 P(node) =0.03622595
## class counts: 3603 3979 1649 664 60
## probabilities: 0.362 0.400 0.166 0.067 0.006
## left son=482 (5563 obs) right son=483 (4392 obs)
## Primary splits:
## depression < 0.5 to the left, improve=24.69099, (0 missing)
## copd < 0.5 to the left, improve=17.49244, (0 missing)
## heart.failure < 0.5 to the left, improve=17.05734, (0 missing)
## reimbursement2008 < 8955 to the left, improve=14.88623, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.99202, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.574, adj=0.034, (0 split)
## age < 47.5 to the right, agree=0.565, adj=0.013, (0 split)
## copd < 0.5 to the left, agree=0.564, adj=0.013, (0 split)
## reimbursement2008 < 13565 to the left, agree=0.561, adj=0.005, (0 split)
## bucket2008 < 3.5 to the left, agree=0.559, adj=0.001, (0 split)
##
## Node number 244: 4305 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.524971 P(node) =0.01566577
## class counts: 1149 2045 746 328 37
## probabilities: 0.267 0.475 0.173 0.076 0.009
## left son=488 (1063 obs) right son=489 (3242 obs)
## Primary splits:
## reimbursement2008 < 9880 to the right, improve=11.346300, (0 missing)
## bucket2008 < 2.5 to the right, improve= 8.562449, (0 missing)
## ihd < 0.5 to the left, improve= 7.353611, (0 missing)
## copd < 0.5 to the left, improve= 6.701463, (0 missing)
## heart.failure < 0.5 to the left, improve= 3.881008, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.941, adj=0.762, (0 split)
##
## Node number 245: 829 observations
## predicted class=B2 expected loss=0.4885404 P(node) =0.003016707
## class counts: 128 424 190 84 3
## probabilities: 0.154 0.511 0.229 0.101 0.004
##
## Node number 248: 7786 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.6050604 P(node) =0.02833302
## class counts: 1982 3075 1667 929 133
## probabilities: 0.255 0.395 0.214 0.119 0.017
## left son=496 (964 obs) right son=497 (6822 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=18.914850, (0 missing)
## depression < 0.5 to the left, improve=16.457650, (0 missing)
## reimbursement2008 < 6325 to the left, improve=12.927220, (0 missing)
## osteoporosis < 0.5 to the left, improve= 9.344273, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.314433, (0 missing)
##
## Node number 249: 1638 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.005960634
## class counts: 210 702 472 227 27
## probabilities: 0.128 0.429 0.288 0.139 0.016
##
## Node number 252: 3345 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.580568 P(node) =0.01217236
## class counts: 372 1403 837 632 101
## probabilities: 0.111 0.419 0.250 0.189 0.030
## left son=504 (1291 obs) right son=505 (2054 obs)
## Primary splits:
## depression < 0.5 to the left, improve=6.733363, (0 missing)
## copd < 0.5 to the left, improve=6.399894, (0 missing)
## cancer < 0.5 to the left, improve=5.398776, (0 missing)
## heart.failure < 0.5 to the left, improve=3.401421, (0 missing)
## age < 31.5 to the right, improve=3.041832, (0 missing)
## Surrogate splits:
## reimbursement2008 < 15665 to the left, agree=0.614, adj=0.001, (0 split)
##
## Node number 253: 2057 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6212931 P(node) =0.007485362
## class counts: 137 779 473 554 114
## probabilities: 0.067 0.379 0.230 0.269 0.055
## left son=506 (520 obs) right son=507 (1537 obs)
## Primary splits:
## copd < 0.5 to the left, improve=4.741452, (0 missing)
## age < 62.5 to the right, improve=3.709690, (0 missing)
## cancer < 0.5 to the right, improve=3.631891, (0 missing)
## ihd < 0.5 to the left, improve=3.269099, (0 missing)
## heart.failure < 0.5 to the left, improve=3.168350, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the left, agree=0.751, adj=0.015, (0 split)
## age < 29 to the left, agree=0.749, adj=0.008, (0 split)
## ihd < 0.5 to the left, agree=0.749, adj=0.008, (0 split)
##
## Node number 254: 5298 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.689128 P(node) =0.01927927
## class counts: 908 1647 1051 1364 328
## probabilities: 0.171 0.311 0.198 0.257 0.062
## left son=508 (2489 obs) right son=509 (2809 obs)
## Primary splits:
## depression < 0.5 to the left, improve=18.17296, (0 missing)
## reimbursement2008 < 22335 to the left, improve=13.06444, (0 missing)
## copd < 0.5 to the left, improve=11.53148, (0 missing)
## ihd < 0.5 to the left, improve= 8.63716, (0 missing)
## heart.failure < 0.5 to the left, improve= 8.42218, (0 missing)
## Surrogate splits:
## copd < 0.5 to the left, agree=0.579, adj=0.104, (0 split)
## alzheimers < 0.5 to the left, agree=0.573, adj=0.092, (0 split)
## ihd < 0.5 to the left, agree=0.545, adj=0.033, (0 split)
## heart.failure < 0.5 to the left, agree=0.544, adj=0.030, (0 split)
## reimbursement2008 < 16955 to the left, agree=0.535, adj=0.010, (0 split)
##
## Node number 255: 1785 observations
## predicted class=B2 expected loss=0.6369748 P(node) =0.006495562
## class counts: 174 648 492 397 74
## probabilities: 0.097 0.363 0.276 0.222 0.041
##
## Node number 406: 128 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5 P(node) =0.0004657882
## class counts: 64 42 14 8 0
## probabilities: 0.500 0.328 0.109 0.062 0.000
## left son=812 (95 obs) right son=813 (33 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.3313600, (0 missing)
## reimbursement2008 < 2155 to the left, improve=2.1875000, (0 missing)
## age < 70.5 to the right, improve=1.5228130, (0 missing)
## arthritis < 0.5 to the left, improve=1.1806970, (0 missing)
## copd < 0.5 to the left, improve=0.4207762, (0 missing)
##
## Node number 407: 158 observations
## predicted class=B2 expected loss=0.5379747 P(node) =0.0005749573
## class counts: 55 73 22 8 0
## probabilities: 0.348 0.462 0.139 0.051 0.000
##
## Node number 412: 270 observations
## predicted class=B1 expected loss=0.462963 P(node) =0.000982522
## class counts: 145 73 36 15 1
## probabilities: 0.537 0.270 0.133 0.056 0.004
##
## Node number 413: 634 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5694006 P(node) =0.002307107
## class counts: 273 231 83 42 5
## probabilities: 0.431 0.364 0.131 0.066 0.008
## left son=826 (596 obs) right son=827 (38 obs)
## Primary splits:
## age < 91.5 to the left, improve=3.6059530, (0 missing)
## ihd < 0.5 to the left, improve=2.2411130, (0 missing)
## reimbursement2008 < 1765 to the right, improve=2.0115470, (0 missing)
## cancer < 0.5 to the left, improve=1.8824720, (0 missing)
## depression < 0.5 to the right, improve=0.5863526, (0 missing)
##
## Node number 422: 763 observations
## predicted class=B1 expected loss=0.5307995 P(node) =0.002776534
## class counts: 358 260 102 40 3
## probabilities: 0.469 0.341 0.134 0.052 0.004
##
## Node number 423: 118 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5338983 P(node) =0.0004293985
## class counts: 41 55 16 6 0
## probabilities: 0.347 0.466 0.136 0.051 0.000
## left son=846 (22 obs) right son=847 (96 obs)
## Primary splits:
## reimbursement2008 < 2865 to the right, improve=2.6611130, (0 missing)
## copd < 0.5 to the left, improve=1.5528850, (0 missing)
## heart.failure < 0.5 to the left, improve=1.3108310, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.2553930, (0 missing)
## age < 89.5 to the left, improve=0.9696791, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=0.873, adj=0.318, (0 split)
##
## Node number 424: 495 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5656566 P(node) =0.00180129
## class counts: 215 202 50 26 2
## probabilities: 0.434 0.408 0.101 0.053 0.004
## left son=848 (385 obs) right son=849 (110 obs)
## Primary splits:
## reimbursement2008 < 2795 to the right, improve=2.2427130, (0 missing)
## age < 73.5 to the left, improve=1.5903260, (0 missing)
## ihd < 0.5 to the left, improve=1.1717170, (0 missing)
## depression < 0.5 to the left, improve=0.5724615, (0 missing)
## kidney < 0.5 to the left, improve=0.5572971, (0 missing)
##
## Node number 425: 943 observations
## predicted class=B1 expected loss=0.5344645 P(node) =0.003431549
## class counts: 439 313 137 50 4
## probabilities: 0.466 0.332 0.145 0.053 0.004
##
## Node number 432: 1265 observations
## predicted class=B1 expected loss=0.4442688 P(node) =0.004603298
## class counts: 703 367 147 44 4
## probabilities: 0.556 0.290 0.116 0.035 0.003
##
## Node number 433: 2727 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5192519 P(node) =0.009923472
## class counts: 1311 948 350 109 9
## probabilities: 0.481 0.348 0.128 0.040 0.003
## left son=866 (1499 obs) right son=867 (1228 obs)
## Primary splits:
## reimbursement2008 < 2615 to the left, improve=4.028460, (0 missing)
## age < 54.5 to the left, improve=3.426946, (0 missing)
## copd < 0.5 to the left, improve=2.215284, (0 missing)
## stroke < 0.5 to the left, improve=2.121876, (0 missing)
## depression < 0.5 to the left, improve=1.918448, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.612, adj=0.139, (0 split)
## age < 97.5 to the left, agree=0.552, adj=0.005, (0 split)
##
## Node number 434: 238 observations, complexity param=0.0001826494
## predicted class=B1 expected loss=0.5714286 P(node) =0.000866075
## class counts: 102 86 38 9 3
## probabilities: 0.429 0.361 0.160 0.038 0.013
## left son=868 (167 obs) right son=869 (71 obs)
## Primary splits:
## depression < 0.5 to the left, improve=4.834875, (0 missing)
## age < 59.5 to the right, improve=2.461134, (0 missing)
## ihd < 0.5 to the left, improve=1.909944, (0 missing)
## reimbursement2008 < 2285 to the left, improve=1.842456, (0 missing)
## alzheimers < 0.5 to the left, improve=1.113912, (0 missing)
## Surrogate splits:
## age < 97.5 to the left, agree=0.718, adj=0.056, (0 split)
##
## Node number 435: 145 observations
## predicted class=B2 expected loss=0.4689655 P(node) =0.0005276507
## class counts: 39 77 20 8 1
## probabilities: 0.269 0.531 0.138 0.055 0.007
##
## Node number 436: 635 observations
## predicted class=B1 expected loss=0.5023622 P(node) =0.002310746
## class counts: 316 196 93 26 4
## probabilities: 0.498 0.309 0.146 0.041 0.006
##
## Node number 437: 2184 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.595696 P(node) =0.007947511
## class counts: 883 784 346 157 14
## probabilities: 0.404 0.359 0.158 0.072 0.006
## left son=874 (393 obs) right son=875 (1791 obs)
## Primary splits:
## reimbursement2008 < 2315 to the left, improve=4.386891, (0 missing)
## depression < 0.5 to the left, improve=4.376862, (0 missing)
## age < 39.5 to the right, improve=3.004733, (0 missing)
## alzheimers < 0.5 to the left, improve=2.391734, (0 missing)
## stroke < 0.5 to the left, improve=2.171601, (0 missing)
##
## Node number 438: 613 observations, complexity param=8.302246e-05
## predicted class=B1 expected loss=0.6182708 P(node) =0.002230689
## class counts: 234 226 111 36 6
## probabilities: 0.382 0.369 0.181 0.059 0.010
## left son=876 (180 obs) right son=877 (433 obs)
## Primary splits:
## osteoporosis < 0.5 to the right, improve=1.4494640, (0 missing)
## age < 98.5 to the right, improve=1.3979840, (0 missing)
## stroke < 0.5 to the left, improve=0.9190213, (0 missing)
## reimbursement2008 < 2275 to the left, improve=0.8284921, (0 missing)
## depression < 0.5 to the left, improve=0.7804891, (0 missing)
## Surrogate splits:
## reimbursement2008 < 2605 to the right, agree=0.713, adj=0.022, (0 split)
## age < 99.5 to the right, agree=0.711, adj=0.017, (0 split)
##
## Node number 439: 606 observations
## predicted class=B2 expected loss=0.5726073 P(node) =0.002205216
## class counts: 172 259 120 49 6
## probabilities: 0.284 0.427 0.198 0.081 0.010
##
## Node number 440: 143 observations
## predicted class=B1 expected loss=0.3986014 P(node) =0.0005203728
## class counts: 86 37 11 9 0
## probabilities: 0.601 0.259 0.077 0.063 0.000
##
## Node number 441: 374 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5802139 P(node) =0.001360975
## class counts: 157 154 46 16 1
## probabilities: 0.420 0.412 0.123 0.043 0.003
## left son=882 (25 obs) right son=883 (349 obs)
## Primary splits:
## reimbursement2008 < 2315 to the left, improve=4.569334, (0 missing)
## cancer < 0.5 to the left, improve=2.355946, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.361181, (0 missing)
## age < 90.5 to the right, improve=1.103565, (0 missing)
## heart.failure < 0.5 to the left, improve=1.082873, (0 missing)
##
## Node number 442: 18 observations
## predicted class=B1 expected loss=0.2777778 P(node) =6.550147e-05
## class counts: 13 4 0 1 0
## probabilities: 0.722 0.222 0.000 0.056 0.000
##
## Node number 443: 1533 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5942596 P(node) =0.005578542
## class counts: 588 622 223 91 9
## probabilities: 0.384 0.406 0.145 0.059 0.006
## left son=886 (1101 obs) right son=887 (432 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=2.2490350, (0 missing)
## cancer < 0.5 to the left, improve=1.4724050, (0 missing)
## stroke < 0.5 to the left, improve=1.3260620, (0 missing)
## reimbursement2008 < 2435 to the left, improve=1.1404580, (0 missing)
## copd < 0.5 to the left, improve=0.9660973, (0 missing)
##
## Node number 450: 810 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.491358 P(node) =0.002947566
## class counts: 412 257 108 33 0
## probabilities: 0.509 0.317 0.133 0.041 0.000
## left son=900 (117 obs) right son=901 (693 obs)
## Primary splits:
## reimbursement2008 < 11525 to the right, improve=5.1504220, (0 missing)
## bucket2008 < 2.5 to the right, improve=2.8801500, (0 missing)
## age < 34.5 to the left, improve=1.2970260, (0 missing)
## stroke < 0.5 to the left, improve=0.8677994, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5279869, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.928, adj=0.504, (0 split)
##
## Node number 451: 33 observations
## predicted class=B2 expected loss=0.4242424 P(node) =0.000120086
## class counts: 6 19 7 1 0
## probabilities: 0.182 0.576 0.212 0.030 0.000
##
## Node number 452: 3304 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.4757869 P(node) =0.01202316
## class counts: 1732 979 389 183 21
## probabilities: 0.524 0.296 0.118 0.055 0.006
## left son=904 (1626 obs) right son=905 (1678 obs)
## Primary splits:
## reimbursement2008 < 5905 to the right, improve=9.971499, (0 missing)
## bucket2008 < 2.5 to the right, improve=7.851328, (0 missing)
## age < 62.5 to the left, improve=4.441278, (0 missing)
## heart.failure < 0.5 to the left, improve=3.580749, (0 missing)
## kidney < 0.5 to the left, improve=1.765354, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.861, adj=0.717, (0 split)
## kidney < 0.5 to the right, agree=0.637, adj=0.262, (0 split)
## heart.failure < 0.5 to the right, agree=0.608, adj=0.204, (0 split)
## copd < 0.5 to the right, agree=0.590, adj=0.166, (0 split)
## alzheimers < 0.5 to the right, agree=0.558, adj=0.102, (0 split)
##
## Node number 453: 962 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5592516 P(node) =0.00350069
## class counts: 424 364 114 55 5
## probabilities: 0.441 0.378 0.119 0.057 0.005
## left son=906 (857 obs) right son=907 (105 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=3.576264, (0 missing)
## heart.failure < 0.5 to the left, improve=3.114700, (0 missing)
## reimbursement2008 < 59635 to the left, improve=2.145451, (0 missing)
## age < 97.5 to the right, improve=1.742305, (0 missing)
## copd < 0.5 to the left, improve=1.012750, (0 missing)
##
## Node number 454: 1518 observations
## predicted class=B1 expected loss=0.5685112 P(node) =0.005523957
## class counts: 655 520 243 92 8
## probabilities: 0.431 0.343 0.160 0.061 0.005
##
## Node number 455: 706 observations, complexity param=0.0001162314
## predicted class=B2 expected loss=0.6232295 P(node) =0.002569113
## class counts: 233 266 118 77 12
## probabilities: 0.330 0.377 0.167 0.109 0.017
## left son=910 (696 obs) right son=911 (10 obs)
## Primary splits:
## reimbursement2008 < 3155 to the right, improve=3.301177, (0 missing)
## heart.failure < 0.5 to the left, improve=3.232296, (0 missing)
## copd < 0.5 to the left, improve=2.330270, (0 missing)
## alzheimers < 0.5 to the left, improve=1.835216, (0 missing)
## age < 92.5 to the right, improve=1.805094, (0 missing)
##
## Node number 456: 440 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5636364 P(node) =0.001601147
## class counts: 177 192 49 20 2
## probabilities: 0.402 0.436 0.111 0.045 0.005
## left son=912 (58 obs) right son=913 (382 obs)
## Primary splits:
## reimbursement2008 < 3155 to the left, improve=3.3827400, (0 missing)
## age < 80.5 to the right, improve=2.1481460, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6300393, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.5745512, (0 missing)
## depression < 0.5 to the right, improve=0.5547491, (0 missing)
##
## Node number 457: 919 observations
## predicted class=B2 expected loss=0.5505985 P(node) =0.003344214
## class counts: 290 413 154 56 6
## probabilities: 0.316 0.449 0.168 0.061 0.007
##
## Node number 460: 412 observations
## predicted class=B1 expected loss=0.4757282 P(node) =0.001499256
## class counts: 216 120 42 30 4
## probabilities: 0.524 0.291 0.102 0.073 0.010
##
## Node number 461: 302 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6059603 P(node) =0.001098969
## class counts: 102 119 49 31 1
## probabilities: 0.338 0.394 0.162 0.103 0.003
## left son=922 (9 obs) right son=923 (293 obs)
## Primary splits:
## age < 92.5 to the right, improve=2.5766490, (0 missing)
## stroke < 0.5 to the right, improve=1.8961920, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.0608030, (0 missing)
## reimbursement2008 < 32980 to the right, improve=1.0319510, (0 missing)
## kidney < 0.5 to the left, improve=0.7951977, (0 missing)
##
## Node number 462: 282 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.6631206 P(node) =0.00102619
## class counts: 88 95 71 23 5
## probabilities: 0.312 0.337 0.252 0.082 0.018
## left son=924 (220 obs) right son=925 (62 obs)
## Primary splits:
## reimbursement2008 < 27390 to the left, improve=2.933452, (0 missing)
## age < 79.5 to the left, improve=2.171675, (0 missing)
## bucket2008 < 4.5 to the right, improve=1.933271, (0 missing)
## kidney < 0.5 to the left, improve=1.142914, (0 missing)
## ihd < 0.5 to the left, improve=1.125301, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.816, adj=0.161, (0 split)
##
## Node number 463: 162 observations
## predicted class=B2 expected loss=0.5246914 P(node) =0.0005895132
## class counts: 30 77 36 15 4
## probabilities: 0.185 0.475 0.222 0.093 0.025
##
## Node number 472: 159 observations, complexity param=6.167383e-05
## predicted class=B1 expected loss=0.591195 P(node) =0.0005785963
## class counts: 65 51 33 8 2
## probabilities: 0.409 0.321 0.208 0.050 0.013
## left son=944 (76 obs) right son=945 (83 obs)
## Primary splits:
## reimbursement2008 < 11995 to the right, improve=3.4294220, (0 missing)
## age < 65 to the left, improve=1.4674530, (0 missing)
## copd < 0.5 to the left, improve=1.0021090, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7722947, (0 missing)
## bucket2008 < 3.5 to the left, improve=0.4091195, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.673, adj=0.316, (0 split)
## alzheimers < 0.5 to the right, agree=0.591, adj=0.145, (0 split)
## copd < 0.5 to the right, agree=0.572, adj=0.105, (0 split)
## heart.failure < 0.5 to the right, agree=0.560, adj=0.079, (0 split)
## age < 85.5 to the right, agree=0.541, adj=0.039, (0 split)
##
## Node number 473: 586 observations
## predicted class=B2 expected loss=0.6075085 P(node) =0.002132437
## class counts: 167 230 117 64 8
## probabilities: 0.285 0.392 0.200 0.109 0.014
##
## Node number 474: 237 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5738397 P(node) =0.000862436
## class counts: 35 101 72 26 3
## probabilities: 0.148 0.426 0.304 0.110 0.013
## left son=948 (126 obs) right son=949 (111 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.2728500, (0 missing)
## reimbursement2008 < 18275 to the left, improve=1.9530690, (0 missing)
## bucket2008 < 3.5 to the left, improve=1.1622440, (0 missing)
## age < 75.5 to the right, improve=1.0471110, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8993424, (0 missing)
## Surrogate splits:
## age < 86.5 to the left, agree=0.599, adj=0.144, (0 split)
## heart.failure < 0.5 to the left, agree=0.586, adj=0.117, (0 split)
## kidney < 0.5 to the left, agree=0.582, adj=0.108, (0 split)
## depression < 0.5 to the left, agree=0.570, adj=0.081, (0 split)
## stroke < 0.5 to the left, agree=0.565, adj=0.072, (0 split)
##
## Node number 475: 72 observations
## predicted class=B3 expected loss=0.6111111 P(node) =0.0002620059
## class counts: 14 16 28 11 3
## probabilities: 0.194 0.222 0.389 0.153 0.042
##
## Node number 478: 226 observations
## predicted class=B2 expected loss=0.6238938 P(node) =0.0008224073
## class counts: 40 85 74 26 1
## probabilities: 0.177 0.376 0.327 0.115 0.004
##
## Node number 479: 71 observations
## predicted class=B3 expected loss=0.5492958 P(node) =0.0002583669
## class counts: 12 19 32 7 1
## probabilities: 0.169 0.268 0.451 0.099 0.014
##
## Node number 480: 403 observations
## predicted class=B1 expected loss=0.4243176 P(node) =0.001466505
## class counts: 232 86 61 20 4
## probabilities: 0.576 0.213 0.151 0.050 0.010
##
## Node number 481: 2214 observations, complexity param=0.0001439056
## predicted class=B1 expected loss=0.5442638 P(node) =0.008056681
## class counts: 1009 798 290 107 10
## probabilities: 0.456 0.360 0.131 0.048 0.005
## left son=962 (1636 obs) right son=963 (578 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=6.829533, (0 missing)
## depression < 0.5 to the left, improve=5.940363, (0 missing)
## copd < 0.5 to the left, improve=3.544519, (0 missing)
## alzheimers < 0.5 to the left, improve=2.874488, (0 missing)
## age < 57.5 to the left, improve=2.182371, (0 missing)
## Surrogate splits:
## reimbursement2008 < 9185 to the left, agree=0.742, adj=0.01, (0 split)
##
## Node number 482: 5563 observations, complexity param=0.000785946
## predicted class=B1 expected loss=0.6002157 P(node) =0.02024359
## class counts: 2224 2125 853 328 33
## probabilities: 0.400 0.382 0.153 0.059 0.006
## left son=964 (1363 obs) right son=965 (4200 obs)
## Primary splits:
## reimbursement2008 < 8955 to the right, improve=10.881900, (0 missing)
## heart.failure < 0.5 to the left, improve= 9.391652, (0 missing)
## copd < 0.5 to the left, improve= 8.088751, (0 missing)
## bucket2008 < 2.5 to the right, improve= 7.479912, (0 missing)
## age < 31.5 to the left, improve= 2.090877, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.959, adj=0.834, (0 split)
## age < 28.5 to the left, agree=0.755, adj=0.001, (0 split)
##
## Node number 483: 4392 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5778689 P(node) =0.01598236
## class counts: 1379 1854 796 336 27
## probabilities: 0.314 0.422 0.181 0.077 0.006
## left son=966 (2928 obs) right son=967 (1464 obs)
## Primary splits:
## reimbursement2008 < 8325 to the left, improve=7.509791, (0 missing)
## copd < 0.5 to the left, improve=6.714633, (0 missing)
## heart.failure < 0.5 to the left, improve=5.287260, (0 missing)
## bucket2008 < 2.5 to the left, improve=5.126640, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.940785, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.979, adj=0.936, (0 split)
##
## Node number 488: 1063 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5983067 P(node) =0.003868226
## class counts: 336 427 192 95 13
## probabilities: 0.316 0.402 0.181 0.089 0.012
## left son=976 (102 obs) right son=977 (961 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=6.866573, (0 missing)
## heart.failure < 0.5 to the left, improve=4.371067, (0 missing)
## copd < 0.5 to the left, improve=3.836869, (0 missing)
## reimbursement2008 < 18130 to the left, improve=3.632764, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.933358, (0 missing)
##
## Node number 489: 3242 observations
## predicted class=B2 expected loss=0.5009254 P(node) =0.01179754
## class counts: 813 1618 554 233 24
## probabilities: 0.251 0.499 0.171 0.072 0.007
##
## Node number 496: 964 observations, complexity param=0.0001411382
## predicted class=B1 expected loss=0.6307054 P(node) =0.003507968
## class counts: 356 348 167 82 11
## probabilities: 0.369 0.361 0.173 0.085 0.011
## left son=992 (572 obs) right son=993 (392 obs)
## Primary splits:
## depression < 0.5 to the left, improve=5.116701, (0 missing)
## reimbursement2008 < 8255 to the left, improve=4.417859, (0 missing)
## bucket2008 < 2.5 to the left, improve=3.940989, (0 missing)
## heart.failure < 0.5 to the left, improve=3.162605, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.038248, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.601, adj=0.018, (0 split)
## reimbursement2008 < 14855 to the left, agree=0.598, adj=0.010, (0 split)
## copd < 0.5 to the left, agree=0.594, adj=0.003, (0 split)
##
## Node number 497: 6822 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6002639 P(node) =0.02482506
## class counts: 1626 2727 1500 847 122
## probabilities: 0.238 0.400 0.220 0.124 0.018
## left son=994 (3172 obs) right son=995 (3650 obs)
## Primary splits:
## reimbursement2008 < 6325 to the left, improve=11.481700, (0 missing)
## depression < 0.5 to the left, improve=11.166950, (0 missing)
## bucket2008 < 2.5 to the left, improve= 7.602059, (0 missing)
## osteoporosis < 0.5 to the left, improve= 6.404955, (0 missing)
## copd < 0.5 to the left, improve= 5.945024, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.869, adj=0.717, (0 split)
## copd < 0.5 to the left, agree=0.576, adj=0.088, (0 split)
## heart.failure < 0.5 to the left, agree=0.570, adj=0.076, (0 split)
## alzheimers < 0.5 to the left, agree=0.545, adj=0.020, (0 split)
## age < 31.5 to the left, agree=0.536, adj=0.003, (0 split)
##
## Node number 504: 1291 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5530596 P(node) =0.004697911
## class counts: 183 577 280 219 32
## probabilities: 0.142 0.447 0.217 0.170 0.025
## left son=1008 (973 obs) right son=1009 (318 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=5.737334, (0 missing)
## reimbursement2008 < 32055 to the left, improve=1.892346, (0 missing)
## age < 84.5 to the right, improve=1.763761, (0 missing)
## ihd < 0.5 to the left, improve=1.710826, (0 missing)
## alzheimers < 0.5 to the right, improve=1.100420, (0 missing)
## Surrogate splits:
## age < 28.5 to the right, agree=0.755, adj=0.006, (0 split)
##
## Node number 505: 2054 observations
## predicted class=B2 expected loss=0.5978578 P(node) =0.007474445
## class counts: 189 826 557 413 69
## probabilities: 0.092 0.402 0.271 0.201 0.034
##
## Node number 506: 520 observations
## predicted class=B2 expected loss=0.5769231 P(node) =0.001892265
## class counts: 50 220 120 108 22
## probabilities: 0.096 0.423 0.231 0.208 0.042
##
## Node number 507: 1537 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6363045 P(node) =0.005593098
## class counts: 87 559 353 446 92
## probabilities: 0.057 0.364 0.230 0.290 0.060
## left son=1014 (1286 obs) right son=1015 (251 obs)
## Primary splits:
## age < 62.5 to the right, improve=4.292573, (0 missing)
## reimbursement2008 < 43950 to the left, improve=3.358938, (0 missing)
## cancer < 0.5 to the right, improve=2.803709, (0 missing)
## heart.failure < 0.5 to the left, improve=1.956332, (0 missing)
## stroke < 0.5 to the left, improve=1.605851, (0 missing)
##
## Node number 508: 2489 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.7219767 P(node) =0.009057397
## class counts: 541 692 436 670 150
## probabilities: 0.217 0.278 0.175 0.269 0.060
## left son=1016 (1317 obs) right son=1017 (1172 obs)
## Primary splits:
## copd < 0.5 to the right, improve=9.293163, (0 missing)
## reimbursement2008 < 23175 to the left, improve=7.265866, (0 missing)
## ihd < 0.5 to the left, improve=7.177016, (0 missing)
## heart.failure < 0.5 to the left, improve=4.187307, (0 missing)
## bucket2008 < 4.5 to the left, improve=3.684093, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the right, agree=0.575, adj=0.097, (0 split)
## alzheimers < 0.5 to the right, agree=0.556, adj=0.057, (0 split)
## ihd < 0.5 to the right, agree=0.555, adj=0.055, (0 split)
## reimbursement2008 < 27380 to the right, agree=0.544, adj=0.032, (0 split)
## age < 52.5 to the right, agree=0.532, adj=0.005, (0 split)
##
## Node number 509: 2809 observations
## predicted class=B2 expected loss=0.6600214 P(node) =0.01022187
## class counts: 367 955 615 694 178
## probabilities: 0.131 0.340 0.219 0.247 0.063
##
## Node number 812: 95 observations
## predicted class=B1 expected loss=0.4315789 P(node) =0.0003457022
## class counts: 54 25 11 5 0
## probabilities: 0.568 0.263 0.116 0.053 0.000
##
## Node number 813: 33 observations
## predicted class=B2 expected loss=0.4848485 P(node) =0.000120086
## class counts: 10 17 3 3 0
## probabilities: 0.303 0.515 0.091 0.091 0.000
##
## Node number 826: 596 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5536913 P(node) =0.002168826
## class counts: 266 214 77 34 5
## probabilities: 0.446 0.359 0.129 0.057 0.008
## left son=1652 (555 obs) right son=1653 (41 obs)
## Primary splits:
## reimbursement2008 < 1765 to the right, improve=2.4482060, (0 missing)
## age < 81.5 to the right, improve=2.3548750, (0 missing)
## ihd < 0.5 to the left, improve=2.3213280, (0 missing)
## cancer < 0.5 to the left, improve=2.1512770, (0 missing)
## depression < 0.5 to the left, improve=0.5755867, (0 missing)
##
## Node number 827: 38 observations
## predicted class=B2 expected loss=0.5526316 P(node) =0.0001382809
## class counts: 7 17 6 8 0
## probabilities: 0.184 0.447 0.158 0.211 0.000
##
## Node number 846: 22 observations
## predicted class=B1 expected loss=0.4545455 P(node) =8.005735e-05
## class counts: 12 5 4 1 0
## probabilities: 0.545 0.227 0.182 0.045 0.000
##
## Node number 847: 96 observations
## predicted class=B2 expected loss=0.4791667 P(node) =0.0003493412
## class counts: 29 50 12 5 0
## probabilities: 0.302 0.521 0.125 0.052 0.000
##
## Node number 848: 385 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5454545 P(node) =0.001401004
## class counts: 175 146 39 23 2
## probabilities: 0.455 0.379 0.101 0.060 0.005
## left son=1696 (263 obs) right son=1697 (122 obs)
## Primary splits:
## age < 80.5 to the left, improve=2.5496070, (0 missing)
## ihd < 0.5 to the left, improve=1.7465940, (0 missing)
## reimbursement2008 < 3025 to the right, improve=0.9395484, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7038989, (0 missing)
## depression < 0.5 to the left, improve=0.3133237, (0 missing)
##
## Node number 849: 110 observations
## predicted class=B2 expected loss=0.4909091 P(node) =0.0004002868
## class counts: 40 56 11 3 0
## probabilities: 0.364 0.509 0.100 0.027 0.000
##
## Node number 866: 1499 observations
## predicted class=B1 expected loss=0.490994 P(node) =0.005454817
## class counts: 763 492 189 52 3
## probabilities: 0.509 0.328 0.126 0.035 0.002
##
## Node number 867: 1228 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5537459 P(node) =0.004468656
## class counts: 548 456 161 57 6
## probabilities: 0.446 0.371 0.131 0.046 0.005
## left son=1734 (171 obs) right son=1735 (1057 obs)
## Primary splits:
## reimbursement2008 < 2995 to the right, improve=3.399761, (0 missing)
## bucket2008 < 1.5 to the right, improve=3.399761, (0 missing)
## age < 83.5 to the left, improve=2.697325, (0 missing)
## kidney < 0.5 to the left, improve=2.465389, (0 missing)
## depression < 0.5 to the left, improve=1.722272, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=1, adj=1, (0 split)
##
## Node number 868: 167 observations
## predicted class=B1 expected loss=0.502994 P(node) =0.0006077081
## class counts: 83 50 25 7 2
## probabilities: 0.497 0.299 0.150 0.042 0.012
##
## Node number 869: 71 observations
## predicted class=B2 expected loss=0.4929577 P(node) =0.0002583669
## class counts: 19 36 13 2 1
## probabilities: 0.268 0.507 0.183 0.028 0.014
##
## Node number 874: 393 observations
## predicted class=B1 expected loss=0.5139949 P(node) =0.001430115
## class counts: 191 134 46 20 2
## probabilities: 0.486 0.341 0.117 0.051 0.005
##
## Node number 875: 1791 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.6136237 P(node) =0.006517396
## class counts: 692 650 300 137 12
## probabilities: 0.386 0.363 0.168 0.076 0.007
## left son=1750 (1752 obs) right son=1751 (39 obs)
## Primary splits:
## age < 39.5 to the right, improve=3.631907, (0 missing)
## depression < 0.5 to the left, improve=3.598994, (0 missing)
## reimbursement2008 < 2475 to the left, improve=1.828499, (0 missing)
## alzheimers < 0.5 to the left, improve=1.790378, (0 missing)
## stroke < 0.5 to the left, improve=1.609073, (0 missing)
##
## Node number 876: 180 observations, complexity param=8.302246e-05
## predicted class=B1 expected loss=0.5666667 P(node) =0.0006550147
## class counts: 78 68 23 11 0
## probabilities: 0.433 0.378 0.128 0.061 0.000
## left son=1752 (112 obs) right son=1753 (68 obs)
## Primary splits:
## reimbursement2008 < 2455 to the left, improve=3.4787820, (0 missing)
## age < 88.5 to the left, improve=1.3486290, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8074074, (0 missing)
## copd < 0.5 to the left, improve=0.6952328, (0 missing)
## ihd < 0.5 to the right, improve=0.3305556, (0 missing)
##
## Node number 877: 433 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.6351039 P(node) =0.001575674
## class counts: 156 158 88 25 6
## probabilities: 0.360 0.365 0.203 0.058 0.014
## left son=1754 (403 obs) right son=1755 (30 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=1.8988510, (0 missing)
## age < 45.5 to the left, improve=1.3010460, (0 missing)
## reimbursement2008 < 2255 to the left, improve=1.2799960, (0 missing)
## depression < 0.5 to the left, improve=1.2338070, (0 missing)
## cancer < 0.5 to the right, improve=0.6525541, (0 missing)
##
## Node number 882: 25 observations
## predicted class=B2 expected loss=0.24 P(node) =9.097426e-05
## class counts: 6 19 0 0 0
## probabilities: 0.240 0.760 0.000 0.000 0.000
##
## Node number 883: 349 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5673352 P(node) =0.001270001
## class counts: 151 135 46 16 1
## probabilities: 0.433 0.387 0.132 0.046 0.003
## left son=1766 (336 obs) right son=1767 (13 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.2574550, (0 missing)
## age < 69.5 to the left, improve=1.5047970, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.0520090, (0 missing)
## heart.failure < 0.5 to the left, improve=0.8771249, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.8193194, (0 missing)
##
## Node number 886: 1101 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.595822 P(node) =0.004006506
## class counts: 445 444 150 57 5
## probabilities: 0.404 0.403 0.136 0.052 0.005
## left son=1772 (1057 obs) right son=1773 (44 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=2.0669290, (0 missing)
## age < 46.5 to the left, improve=1.0570040, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.9632398, (0 missing)
## cancer < 0.5 to the left, improve=0.9197684, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7998904, (0 missing)
##
## Node number 887: 432 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.587963 P(node) =0.001572035
## class counts: 143 178 73 34 4
## probabilities: 0.331 0.412 0.169 0.079 0.009
## left son=1774 (403 obs) right son=1775 (29 obs)
## Primary splits:
## reimbursement2008 < 2215 to the right, improve=1.979831, (0 missing)
## depression < 0.5 to the left, improve=1.399095, (0 missing)
## age < 65.5 to the right, improve=1.336452, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.190308, (0 missing)
## heart.failure < 0.5 to the left, improve=1.062301, (0 missing)
##
## Node number 900: 117 observations
## predicted class=B1 expected loss=0.3418803 P(node) =0.0004257595
## class counts: 77 25 8 7 0
## probabilities: 0.658 0.214 0.068 0.060 0.000
##
## Node number 901: 693 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5165945 P(node) =0.002521807
## class counts: 335 232 100 26 0
## probabilities: 0.483 0.335 0.144 0.038 0.000
## left son=1802 (684 obs) right son=1803 (9 obs)
## Primary splits:
## reimbursement2008 < 11105 to the left, improve=2.2165640, (0 missing)
## alzheimers < 0.5 to the left, improve=1.2536740, (0 missing)
## stroke < 0.5 to the left, improve=0.8354877, (0 missing)
## age < 34 to the left, improve=0.8168384, (0 missing)
## copd < 0.5 to the left, improve=0.3325841, (0 missing)
##
## Node number 904: 1626 observations
## predicted class=B1 expected loss=0.4391144 P(node) =0.005916966
## class counts: 912 414 190 99 11
## probabilities: 0.561 0.255 0.117 0.061 0.007
##
## Node number 905: 1678 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.511323 P(node) =0.006106192
## class counts: 820 565 199 84 10
## probabilities: 0.489 0.337 0.119 0.050 0.006
## left son=1810 (1608 obs) right son=1811 (70 obs)
## Primary splits:
## reimbursement2008 < 5695 to the left, improve=3.4228850, (0 missing)
## age < 54.5 to the left, improve=3.4011680, (0 missing)
## kidney < 0.5 to the left, improve=2.7930360, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6256038, (0 missing)
## stroke < 0.5 to the left, improve=0.4872841, (0 missing)
##
## Node number 906: 857 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5425904 P(node) =0.003118598
## class counts: 392 313 100 48 4
## probabilities: 0.457 0.365 0.117 0.056 0.005
## left son=1812 (405 obs) right son=1813 (452 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=1.8446560, (0 missing)
## reimbursement2008 < 8165 to the right, improve=1.5511950, (0 missing)
## age < 58.5 to the right, improve=1.3750490, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.0178390, (0 missing)
## kidney < 0.5 to the left, improve=0.9078092, (0 missing)
## Surrogate splits:
## reimbursement2008 < 7010 to the left, agree=0.629, adj=0.215, (0 split)
## bucket2008 < 2.5 to the left, agree=0.611, adj=0.178, (0 split)
## kidney < 0.5 to the left, agree=0.585, adj=0.121, (0 split)
## copd < 0.5 to the left, agree=0.583, adj=0.119, (0 split)
## age < 77.5 to the left, agree=0.555, adj=0.059, (0 split)
##
## Node number 907: 105 observations
## predicted class=B2 expected loss=0.5142857 P(node) =0.0003820919
## class counts: 32 51 14 7 1
## probabilities: 0.305 0.486 0.133 0.067 0.010
##
## Node number 910: 696 observations, complexity param=0.0001162314
## predicted class=B2 expected loss=0.6192529 P(node) =0.002532723
## class counts: 232 265 112 75 12
## probabilities: 0.333 0.381 0.161 0.108 0.017
## left son=1820 (177 obs) right son=1821 (519 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.566495, (0 missing)
## copd < 0.5 to the left, improve=2.688595, (0 missing)
## age < 70.5 to the right, improve=1.873250, (0 missing)
## alzheimers < 0.5 to the left, improve=1.737201, (0 missing)
## reimbursement2008 < 7405 to the left, improve=1.699902, (0 missing)
##
## Node number 911: 10 observations
## predicted class=B3 expected loss=0.4 P(node) =3.63897e-05
## class counts: 1 1 6 2 0
## probabilities: 0.100 0.100 0.600 0.200 0.000
##
## Node number 912: 58 observations
## predicted class=B2 expected loss=0.3793103 P(node) =0.0002110603
## class counts: 20 36 0 1 1
## probabilities: 0.345 0.621 0.000 0.017 0.017
##
## Node number 913: 382 observations, complexity param=5.313437e-05
## predicted class=B1 expected loss=0.5890052 P(node) =0.001390087
## class counts: 157 156 49 19 1
## probabilities: 0.411 0.408 0.128 0.050 0.003
## left son=1826 (25 obs) right son=1827 (357 obs)
## Primary splits:
## reimbursement2008 < 3245 to the left, improve=2.6514540, (0 missing)
## age < 80.5 to the right, improve=2.1655330, (0 missing)
## depression < 0.5 to the left, improve=1.2095670, (0 missing)
## kidney < 0.5 to the left, improve=1.1024610, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7876318, (0 missing)
##
## Node number 922: 9 observations
## predicted class=B1 expected loss=0.3333333 P(node) =3.275073e-05
## class counts: 6 0 1 1 1
## probabilities: 0.667 0.000 0.111 0.111 0.111
##
## Node number 923: 293 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5938567 P(node) =0.001066218
## class counts: 96 119 48 30 0
## probabilities: 0.328 0.406 0.164 0.102 0.000
## left son=1846 (39 obs) right son=1847 (254 obs)
## Primary splits:
## stroke < 0.5 to the right, improve=2.1766940, (0 missing)
## age < 65.5 to the right, improve=1.5636490, (0 missing)
## reimbursement2008 < 11660 to the right, improve=1.0372370, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9901623, (0 missing)
## kidney < 0.5 to the left, improve=0.7188410, (0 missing)
## Surrogate splits:
## reimbursement2008 < 79760 to the right, agree=0.87, adj=0.026, (0 split)
##
## Node number 924: 220 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.65 P(node) =0.0008005735
## class counts: 77 66 57 16 4
## probabilities: 0.350 0.300 0.259 0.073 0.018
## left son=1848 (132 obs) right son=1849 (88 obs)
## Primary splits:
## reimbursement2008 < 12810 to the right, improve=2.5787880, (0 missing)
## age < 76.5 to the left, improve=2.1718510, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.0384950, (0 missing)
## stroke < 0.5 to the left, improve=0.8392769, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7969697, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.636, adj=0.091, (0 split)
##
## Node number 925: 62 observations
## predicted class=B2 expected loss=0.5322581 P(node) =0.0002256162
## class counts: 11 29 14 7 1
## probabilities: 0.177 0.468 0.226 0.113 0.016
##
## Node number 944: 76 observations
## predicted class=B1 expected loss=0.4736842 P(node) =0.0002765618
## class counts: 40 18 12 5 1
## probabilities: 0.526 0.237 0.158 0.066 0.013
##
## Node number 945: 83 observations
## predicted class=B2 expected loss=0.6024096 P(node) =0.0003020345
## class counts: 25 33 21 3 1
## probabilities: 0.301 0.398 0.253 0.036 0.012
##
## Node number 948: 126 observations
## predicted class=B2 expected loss=0.5079365 P(node) =0.0004585103
## class counts: 22 62 33 9 0
## probabilities: 0.175 0.492 0.262 0.071 0.000
##
## Node number 949: 111 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6486486 P(node) =0.0004039257
## class counts: 13 39 39 17 3
## probabilities: 0.117 0.351 0.351 0.153 0.027
## left son=1898 (54 obs) right son=1899 (57 obs)
## Primary splits:
## age < 75.5 to the left, improve=1.9702330, (0 missing)
## reimbursement2008 < 23940 to the left, improve=1.5183950, (0 missing)
## stroke < 0.5 to the right, improve=1.4096100, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.1218360, (0 missing)
## kidney < 0.5 to the right, improve=0.9749865, (0 missing)
## Surrogate splits:
## reimbursement2008 < 36125 to the right, agree=0.586, adj=0.148, (0 split)
## depression < 0.5 to the right, agree=0.568, adj=0.111, (0 split)
## alzheimers < 0.5 to the right, agree=0.550, adj=0.074, (0 split)
## bucket2008 < 4.5 to the right, agree=0.532, adj=0.037, (0 split)
##
## Node number 962: 1636 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5177262 P(node) =0.005953356
## class counts: 789 562 198 80 7
## probabilities: 0.482 0.344 0.121 0.049 0.004
## left son=1924 (1127 obs) right son=1925 (509 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.275005, (0 missing)
## depression < 0.5 to the left, improve=2.200834, (0 missing)
## age < 56.5 to the right, improve=2.161392, (0 missing)
## reimbursement2008 < 3635 to the left, improve=1.571205, (0 missing)
## copd < 0.5 to the left, improve=1.483908, (0 missing)
## Surrogate splits:
## reimbursement2008 < 9210 to the left, agree=0.69, adj=0.004, (0 split)
##
## Node number 963: 578 observations, complexity param=0.0001439056
## predicted class=B2 expected loss=0.5916955 P(node) =0.002103325
## class counts: 220 236 92 27 3
## probabilities: 0.381 0.408 0.159 0.047 0.005
## left son=1926 (339 obs) right son=1927 (239 obs)
## Primary splits:
## depression < 0.5 to the left, improve=4.643194, (0 missing)
## copd < 0.5 to the left, improve=3.299852, (0 missing)
## reimbursement2008 < 4535 to the left, improve=1.771265, (0 missing)
## alzheimers < 0.5 to the left, improve=1.392171, (0 missing)
## age < 39.5 to the right, improve=1.271983, (0 missing)
## Surrogate splits:
## age < 43.5 to the right, agree=0.602, adj=0.038, (0 split)
##
## Node number 964: 1363 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5561262 P(node) =0.004959917
## class counts: 605 435 219 92 12
## probabilities: 0.444 0.319 0.161 0.067 0.009
## left son=1928 (798 obs) right son=1929 (565 obs)
## Primary splits:
## copd < 0.5 to the left, improve=7.963265, (0 missing)
## heart.failure < 0.5 to the left, improve=4.055475, (0 missing)
## reimbursement2008 < 29005 to the left, improve=3.506334, (0 missing)
## stroke < 0.5 to the left, improve=1.818093, (0 missing)
## alzheimers < 0.5 to the left, improve=1.815648, (0 missing)
## Surrogate splits:
## reimbursement2008 < 55265 to the left, agree=0.589, adj=0.009, (0 split)
## bucket2008 < 4.5 to the left, agree=0.589, adj=0.009, (0 split)
## age < 27.5 to the right, agree=0.588, adj=0.005, (0 split)
##
## Node number 965: 4200 observations, complexity param=0.0004649258
## predicted class=B2 expected loss=0.597619 P(node) =0.01528368
## class counts: 1619 1690 634 236 21
## probabilities: 0.385 0.402 0.151 0.056 0.005
## left son=1930 (1953 obs) right son=1931 (2247 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=7.153470, (0 missing)
## copd < 0.5 to the left, improve=3.796544, (0 missing)
## age < 49.5 to the left, improve=2.763159, (0 missing)
## reimbursement2008 < 3415 to the left, improve=2.562311, (0 missing)
## alzheimers < 0.5 to the left, improve=1.867356, (0 missing)
## Surrogate splits:
## reimbursement2008 < 4495 to the left, agree=0.563, adj=0.060, (0 split)
## copd < 0.5 to the left, agree=0.551, adj=0.035, (0 split)
## alzheimers < 0.5 to the left, agree=0.540, adj=0.010, (0 split)
## age < 45.5 to the left, agree=0.536, adj=0.002, (0 split)
##
## Node number 966: 2928 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5529372 P(node) =0.01065491
## class counts: 904 1309 506 190 19
## probabilities: 0.309 0.447 0.173 0.065 0.006
## left son=1932 (1987 obs) right son=1933 (941 obs)
## Primary splits:
## copd < 0.5 to the left, improve=5.088653, (0 missing)
## osteoporosis < 0.5 to the left, improve=4.205973, (0 missing)
## reimbursement2008 < 8045 to the left, improve=3.909055, (0 missing)
## heart.failure < 0.5 to the left, improve=3.356901, (0 missing)
## stroke < 0.5 to the left, improve=3.071040, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8235 to the left, agree=0.68, adj=0.005, (0 split)
##
## Node number 967: 1464 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6277322 P(node) =0.005327453
## class counts: 475 545 290 146 8
## probabilities: 0.324 0.372 0.198 0.100 0.005
## left son=1934 (36 obs) right son=1935 (1428 obs)
## Primary splits:
## reimbursement2008 < 8485 to the left, improve=3.191750, (0 missing)
## age < 78.5 to the left, improve=2.281932, (0 missing)
## heart.failure < 0.5 to the left, improve=2.180745, (0 missing)
## stroke < 0.5 to the left, improve=1.944689, (0 missing)
## copd < 0.5 to the left, improve=1.512341, (0 missing)
##
## Node number 976: 102 observations
## predicted class=B1 expected loss=0.4803922 P(node) =0.000371175
## class counts: 53 28 13 7 1
## probabilities: 0.520 0.275 0.127 0.069 0.010
##
## Node number 977: 961 observations
## predicted class=B2 expected loss=0.5848075 P(node) =0.003497051
## class counts: 283 399 179 88 12
## probabilities: 0.294 0.415 0.186 0.092 0.012
##
## Node number 992: 572 observations, complexity param=0.0001411382
## predicted class=B1 expected loss=0.5909091 P(node) =0.002081491
## class counts: 234 183 92 57 6
## probabilities: 0.409 0.320 0.161 0.100 0.010
## left son=1984 (101 obs) right son=1985 (471 obs)
## Primary splits:
## reimbursement2008 < 3545 to the left, improve=4.251847, (0 missing)
## bucket2008 < 2.5 to the left, improve=2.618377, (0 missing)
## age < 69.5 to the right, improve=2.566628, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.664728, (0 missing)
## heart.failure < 0.5 to the left, improve=1.532473, (0 missing)
##
## Node number 993: 392 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5790816 P(node) =0.001426476
## class counts: 122 165 75 25 5
## probabilities: 0.311 0.421 0.191 0.064 0.013
## left son=1986 (9 obs) right son=1987 (383 obs)
## Primary splits:
## reimbursement2008 < 14460 to the right, improve=2.744913, (0 missing)
## age < 48.5 to the left, improve=1.556717, (0 missing)
## copd < 0.5 to the left, improve=1.522824, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.319075, (0 missing)
## heart.failure < 0.5 to the left, improve=1.292043, (0 missing)
##
## Node number 994: 3172 observations
## predicted class=B2 expected loss=0.5630517 P(node) =0.01154281
## class counts: 709 1386 688 337 52
## probabilities: 0.224 0.437 0.217 0.106 0.016
##
## Node number 995: 3650 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6326027 P(node) =0.01328224
## class counts: 917 1341 812 510 70
## probabilities: 0.251 0.367 0.222 0.140 0.019
## left son=1990 (2424 obs) right son=1991 (1226 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=8.187206, (0 missing)
## depression < 0.5 to the left, improve=5.387571, (0 missing)
## copd < 0.5 to the left, improve=5.189054, (0 missing)
## alzheimers < 0.5 to the left, improve=3.710849, (0 missing)
## heart.failure < 0.5 to the left, improve=2.971629, (0 missing)
##
## Node number 1008: 973 observations
## predicted class=B2 expected loss=0.5611511 P(node) =0.003540718
## class counts: 160 427 184 174 28
## probabilities: 0.164 0.439 0.189 0.179 0.029
##
## Node number 1009: 318 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5283019 P(node) =0.001157193
## class counts: 23 150 96 45 4
## probabilities: 0.072 0.472 0.302 0.142 0.013
## left son=2018 (293 obs) right son=2019 (25 obs)
## Primary splits:
## reimbursement2008 < 16525 to the right, improve=7.797134, (0 missing)
## ihd < 0.5 to the left, improve=3.194748, (0 missing)
## heart.failure < 0.5 to the left, improve=1.170376, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.159119, (0 missing)
## age < 55 to the right, improve=1.113448, (0 missing)
##
## Node number 1014: 1286 observations
## predicted class=B2 expected loss=0.6251944 P(node) =0.004679716
## class counts: 74 482 303 348 79
## probabilities: 0.058 0.375 0.236 0.271 0.061
##
## Node number 1015: 251 observations, complexity param=6.088314e-05
## predicted class=B4 expected loss=0.6095618 P(node) =0.0009133816
## class counts: 13 77 50 98 13
## probabilities: 0.052 0.307 0.199 0.390 0.052
## left son=2030 (237 obs) right son=2031 (14 obs)
## Primary splits:
## reimbursement2008 < 101585 to the left, improve=3.425401, (0 missing)
## age < 61.5 to the left, improve=2.440583, (0 missing)
## heart.failure < 0.5 to the left, improve=2.158559, (0 missing)
## alzheimers < 0.5 to the left, improve=2.020557, (0 missing)
## cancer < 0.5 to the right, improve=1.778561, (0 missing)
##
## Node number 1016: 1317 observations, complexity param=0.0001439056
## predicted class=B2 expected loss=0.6757783 P(node) =0.004792524
## class counts: 269 427 234 313 74
## probabilities: 0.204 0.324 0.178 0.238 0.056
## left son=2032 (72 obs) right son=2033 (1245 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=5.587185, (0 missing)
## reimbursement2008 < 22435 to the left, improve=5.039175, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.150989, (0 missing)
## age < 50.5 to the right, improve=2.709964, (0 missing)
## heart.failure < 0.5 to the left, improve=2.161876, (0 missing)
##
## Node number 1017: 1172 observations, complexity param=0.0002036818
## predicted class=B4 expected loss=0.6953925 P(node) =0.004264873
## class counts: 272 265 202 357 76
## probabilities: 0.232 0.226 0.172 0.305 0.065
## left son=2034 (191 obs) right son=2035 (981 obs)
## Primary splits:
## reimbursement2008 < 43640 to the right, improve=6.105110, (0 missing)
## bucket2008 < 4.5 to the right, improve=4.295055, (0 missing)
## ihd < 0.5 to the left, improve=2.740224, (0 missing)
## heart.failure < 0.5 to the left, improve=2.395917, (0 missing)
## alzheimers < 0.5 to the right, improve=1.864237, (0 missing)
## Surrogate splits:
## bucket2008 < 4.5 to the right, agree=0.925, adj=0.539, (0 split)
##
## Node number 1652: 555 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5441441 P(node) =0.002019629
## class counts: 253 193 76 29 4
## probabilities: 0.456 0.348 0.137 0.052 0.007
## left son=3304 (265 obs) right son=3305 (290 obs)
## Primary splits:
## reimbursement2008 < 1955 to the left, improve=2.2492970, (0 missing)
## age < 44.5 to the right, improve=2.2010530, (0 missing)
## cancer < 0.5 to the left, improve=2.0271900, (0 missing)
## ihd < 0.5 to the left, improve=2.0227740, (0 missing)
## stroke < 0.5 to the left, improve=0.6965966, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.539, adj=0.034, (0 split)
## age < 90.5 to the right, agree=0.532, adj=0.019, (0 split)
## cancer < 0.5 to the right, agree=0.528, adj=0.011, (0 split)
##
## Node number 1653: 41 observations
## predicted class=B2 expected loss=0.4878049 P(node) =0.0001491978
## class counts: 13 21 1 5 1
## probabilities: 0.317 0.512 0.024 0.122 0.024
##
## Node number 1696: 263 observations
## predicted class=B1 expected loss=0.4980989 P(node) =0.0009570492
## class counts: 132 94 26 11 0
## probabilities: 0.502 0.357 0.099 0.042 0.000
##
## Node number 1697: 122 observations
## predicted class=B2 expected loss=0.5737705 P(node) =0.0004439544
## class counts: 43 52 13 12 2
## probabilities: 0.352 0.426 0.107 0.098 0.016
##
## Node number 1734: 171 observations
## predicted class=B1 expected loss=0.4736842 P(node) =0.0006222639
## class counts: 90 46 24 10 1
## probabilities: 0.526 0.269 0.140 0.058 0.006
##
## Node number 1735: 1057 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5666982 P(node) =0.003846392
## class counts: 458 410 137 47 5
## probabilities: 0.433 0.388 0.130 0.044 0.005
## left son=3470 (840 obs) right son=3471 (217 obs)
## Primary splits:
## age < 83.5 to the left, improve=3.809819, (0 missing)
## kidney < 0.5 to the left, improve=2.564065, (0 missing)
## depression < 0.5 to the left, improve=1.351420, (0 missing)
## copd < 0.5 to the left, improve=1.145117, (0 missing)
## reimbursement2008 < 2975 to the left, improve=1.026292, (0 missing)
##
## Node number 1750: 1752 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.6084475 P(node) =0.006375476
## class counts: 686 629 294 131 12
## probabilities: 0.392 0.359 0.168 0.075 0.007
## left son=3500 (1099 obs) right son=3501 (653 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.008256, (0 missing)
## age < 97.5 to the left, improve=2.167182, (0 missing)
## reimbursement2008 < 3055 to the left, improve=1.739250, (0 missing)
## alzheimers < 0.5 to the left, improve=1.536121, (0 missing)
## stroke < 0.5 to the left, improve=1.306511, (0 missing)
## Surrogate splits:
## age < 41.5 to the right, agree=0.628, adj=0.002, (0 split)
##
## Node number 1751: 39 observations
## predicted class=B2 expected loss=0.4615385 P(node) =0.0001419198
## class counts: 6 21 6 6 0
## probabilities: 0.154 0.538 0.154 0.154 0.000
##
## Node number 1752: 112 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0004075647
## class counts: 56 33 14 9 0
## probabilities: 0.500 0.295 0.125 0.080 0.000
##
## Node number 1753: 68 observations
## predicted class=B2 expected loss=0.4852941 P(node) =0.00024745
## class counts: 22 35 9 2 0
## probabilities: 0.324 0.515 0.132 0.029 0.000
##
## Node number 1754: 403 observations, complexity param=6.272808e-05
## predicted class=B1 expected loss=0.6253102 P(node) =0.001466505
## class counts: 151 147 79 20 6
## probabilities: 0.375 0.365 0.196 0.050 0.015
## left son=3508 (382 obs) right son=3509 (21 obs)
## Primary splits:
## reimbursement2008 < 2585 to the left, improve=1.3835870, (0 missing)
## age < 45.5 to the left, improve=1.1912610, (0 missing)
## depression < 0.5 to the left, improve=0.8974996, (0 missing)
## cancer < 0.5 to the right, improve=0.7248908, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2961750, (0 missing)
##
## Node number 1755: 30 observations
## predicted class=B2 expected loss=0.6333333 P(node) =0.0001091691
## class counts: 5 11 9 5 0
## probabilities: 0.167 0.367 0.300 0.167 0.000
##
## Node number 1766: 336 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5535714 P(node) =0.001222694
## class counts: 150 128 44 14 0
## probabilities: 0.446 0.381 0.131 0.042 0.000
## left son=3532 (322 obs) right son=3533 (14 obs)
## Primary splits:
## age < 90.5 to the left, improve=1.4565220, (0 missing)
## heart.failure < 0.5 to the left, improve=1.1063780, (0 missing)
## reimbursement2008 < 2325 to the right, improve=0.8683190, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8476676, (0 missing)
## stroke < 0.5 to the left, improve=0.4098214, (0 missing)
##
## Node number 1767: 13 observations
## predicted class=B2 expected loss=0.4615385 P(node) =4.730662e-05
## class counts: 1 7 2 2 1
## probabilities: 0.077 0.538 0.154 0.154 0.077
##
## Node number 1772: 1057 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.589404 P(node) =0.003846392
## class counts: 434 420 145 54 4
## probabilities: 0.411 0.397 0.137 0.051 0.004
## left son=3544 (1008 obs) right son=3545 (49 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=1.0144070, (0 missing)
## age < 46.5 to the left, improve=1.0088960, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.9481247, (0 missing)
## copd < 0.5 to the left, improve=0.6576908, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5672579, (0 missing)
##
## Node number 1773: 44 observations
## predicted class=B2 expected loss=0.4545455 P(node) =0.0001601147
## class counts: 11 24 5 3 1
## probabilities: 0.250 0.545 0.114 0.068 0.023
##
## Node number 1774: 403 observations
## predicted class=B2 expected loss=0.5756824 P(node) =0.001466505
## class counts: 130 171 70 28 4
## probabilities: 0.323 0.424 0.174 0.069 0.010
##
## Node number 1775: 29 observations
## predicted class=B1 expected loss=0.5517241 P(node) =0.0001055301
## class counts: 13 7 3 6 0
## probabilities: 0.448 0.241 0.103 0.207 0.000
##
## Node number 1802: 684 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5146199 P(node) =0.002489056
## class counts: 332 231 95 26 0
## probabilities: 0.485 0.338 0.139 0.038 0.000
## left son=3604 (286 obs) right son=3605 (398 obs)
## Primary splits:
## reimbursement2008 < 4365 to the left, improve=1.4254390, (0 missing)
## alzheimers < 0.5 to the left, improve=1.1902870, (0 missing)
## stroke < 0.5 to the left, improve=0.8341619, (0 missing)
## age < 34 to the left, improve=0.8175360, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3590913, (0 missing)
##
## Node number 1803: 9 observations
## predicted class=B3 expected loss=0.4444444 P(node) =3.275073e-05
## class counts: 3 1 5 0 0
## probabilities: 0.333 0.111 0.556 0.000 0.000
##
## Node number 1810: 1608 observations
## predicted class=B1 expected loss=0.5062189 P(node) =0.005851465
## class counts: 794 529 193 82 10
## probabilities: 0.494 0.329 0.120 0.051 0.006
##
## Node number 1811: 70 observations
## predicted class=B2 expected loss=0.4857143 P(node) =0.0002547279
## class counts: 26 36 6 2 0
## probabilities: 0.371 0.514 0.086 0.029 0.000
##
## Node number 1812: 405 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5012346 P(node) =0.001473783
## class counts: 202 140 42 18 3
## probabilities: 0.499 0.346 0.104 0.044 0.007
## left son=3624 (329 obs) right son=3625 (76 obs)
## Primary splits:
## age < 83.5 to the left, improve=1.8474760, (0 missing)
## reimbursement2008 < 14045 to the left, improve=1.4437850, (0 missing)
## kidney < 0.5 to the right, improve=1.0197570, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4573240, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4260458, (0 missing)
##
## Node number 1813: 452 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.579646 P(node) =0.001644815
## class counts: 190 173 58 30 1
## probabilities: 0.420 0.383 0.128 0.066 0.002
## left son=3626 (362 obs) right son=3627 (90 obs)
## Primary splits:
## reimbursement2008 < 3875 to the right, improve=2.645100, (0 missing)
## age < 84.5 to the left, improve=2.429876, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.612268, (0 missing)
## alzheimers < 0.5 to the right, improve=1.063100, (0 missing)
## kidney < 0.5 to the left, improve=0.663279, (0 missing)
## Surrogate splits:
## age < 32 to the right, agree=0.803, adj=0.011, (0 split)
##
## Node number 1820: 177 observations
## predicted class=B1 expected loss=0.559322 P(node) =0.0006440978
## class counts: 78 62 26 11 0
## probabilities: 0.441 0.350 0.147 0.062 0.000
##
## Node number 1821: 519 observations
## predicted class=B2 expected loss=0.6088632 P(node) =0.001888626
## class counts: 154 203 86 64 12
## probabilities: 0.297 0.391 0.166 0.123 0.023
##
## Node number 1826: 25 observations
## predicted class=B1 expected loss=0.32 P(node) =9.097426e-05
## class counts: 17 7 1 0 0
## probabilities: 0.680 0.280 0.040 0.000 0.000
##
## Node number 1827: 357 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5826331 P(node) =0.001299112
## class counts: 140 149 48 19 1
## probabilities: 0.392 0.417 0.134 0.053 0.003
## left son=3654 (91 obs) right son=3655 (266 obs)
## Primary splits:
## age < 80.5 to the right, improve=2.1730570, (0 missing)
## reimbursement2008 < 4405 to the right, improve=2.0106540, (0 missing)
## depression < 0.5 to the left, improve=0.7793758, (0 missing)
## kidney < 0.5 to the left, improve=0.7738464, (0 missing)
## alzheimers < 0.5 to the right, improve=0.6298514, (0 missing)
##
## Node number 1846: 39 observations
## predicted class=B1 expected loss=0.4871795 P(node) =0.0001419198
## class counts: 20 12 4 3 0
## probabilities: 0.513 0.308 0.103 0.077 0.000
##
## Node number 1847: 254 observations
## predicted class=B2 expected loss=0.5787402 P(node) =0.0009242985
## class counts: 76 107 44 27 0
## probabilities: 0.299 0.421 0.173 0.106 0.000
##
## Node number 1848: 132 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5909091 P(node) =0.0004803441
## class counts: 54 42 26 9 1
## probabilities: 0.409 0.318 0.197 0.068 0.008
## left son=3696 (105 obs) right son=3697 (27 obs)
## Primary splits:
## age < 84.5 to the left, improve=4.2569990, (0 missing)
## reimbursement2008 < 13440 to the left, improve=1.5425260, (0 missing)
## ihd < 0.5 to the right, improve=1.3693110, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4363743, (0 missing)
## kidney < 0.5 to the left, improve=0.4353832, (0 missing)
##
## Node number 1849: 88 observations
## predicted class=B3 expected loss=0.6477273 P(node) =0.0003202294
## class counts: 23 24 31 7 3
## probabilities: 0.261 0.273 0.352 0.080 0.034
##
## Node number 1898: 54 observations
## predicted class=B3 expected loss=0.5555556 P(node) =0.0001965044
## class counts: 8 14 24 7 1
## probabilities: 0.148 0.259 0.444 0.130 0.019
##
## Node number 1899: 57 observations
## predicted class=B2 expected loss=0.5614035 P(node) =0.0002074213
## class counts: 5 25 15 10 2
## probabilities: 0.088 0.439 0.263 0.175 0.035
##
## Node number 1924: 1127 observations
## predicted class=B1 expected loss=0.4960071 P(node) =0.00410112
## class counts: 568 377 131 47 4
## probabilities: 0.504 0.335 0.116 0.042 0.004
##
## Node number 1925: 509 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5658153 P(node) =0.001852236
## class counts: 221 185 67 33 3
## probabilities: 0.434 0.363 0.132 0.065 0.006
## left son=3850 (137 obs) right son=3851 (372 obs)
## Primary splits:
## reimbursement2008 < 3775 to the left, improve=1.6880360, (0 missing)
## depression < 0.5 to the left, improve=1.6361880, (0 missing)
## age < 96.5 to the left, improve=1.5026800, (0 missing)
## copd < 0.5 to the left, improve=1.2566690, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7418596, (0 missing)
##
## Node number 1926: 339 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5575221 P(node) =0.001233611
## class counts: 150 127 45 16 1
## probabilities: 0.442 0.375 0.133 0.047 0.003
## left son=3852 (211 obs) right son=3853 (128 obs)
## Primary splits:
## reimbursement2008 < 4905 to the left, improve=1.6963240, (0 missing)
## age < 45 to the right, improve=1.4829560, (0 missing)
## heart.failure < 0.5 to the left, improve=1.2573130, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6055009, (0 missing)
## copd < 0.5 to the left, improve=0.2708942, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.687, adj=0.172, (0 split)
## copd < 0.5 to the left, agree=0.664, adj=0.109, (0 split)
## stroke < 0.5 to the left, agree=0.652, adj=0.078, (0 split)
##
## Node number 1927: 239 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.5439331 P(node) =0.0008697139
## class counts: 70 109 47 11 2
## probabilities: 0.293 0.456 0.197 0.046 0.008
## left son=3854 (181 obs) right son=3855 (58 obs)
## Primary splits:
## copd < 0.5 to the left, improve=6.3468870, (0 missing)
## reimbursement2008 < 5790 to the right, improve=1.7891020, (0 missing)
## age < 60.5 to the left, improve=1.2691270, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8740764, (0 missing)
## stroke < 0.5 to the right, improve=0.6684821, (0 missing)
## Surrogate splits:
## age < 35 to the right, agree=0.762, adj=0.017, (0 split)
##
## Node number 1928: 798 observations
## predicted class=B1 expected loss=0.5075188 P(node) =0.002903898
## class counts: 393 256 95 51 3
## probabilities: 0.492 0.321 0.119 0.064 0.004
##
## Node number 1929: 565 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.6247788 P(node) =0.002056018
## class counts: 212 179 124 41 9
## probabilities: 0.375 0.317 0.219 0.073 0.016
## left son=3858 (116 obs) right son=3859 (449 obs)
## Primary splits:
## stroke < 0.5 to the right, improve=4.0381830, (0 missing)
## reimbursement2008 < 31655 to the left, improve=2.7523450, (0 missing)
## age < 42.5 to the right, improve=1.7655450, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.4692280, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5615109, (0 missing)
## Surrogate splits:
## reimbursement2008 < 61780 to the right, agree=0.8, adj=0.026, (0 split)
##
## Node number 1930: 1953 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.578085 P(node) =0.007106909
## class counts: 824 782 251 88 8
## probabilities: 0.422 0.400 0.129 0.045 0.004
## left son=3860 (343 obs) right son=3861 (1610 obs)
## Primary splits:
## reimbursement2008 < 3415 to the left, improve=3.4037160, (0 missing)
## age < 42.5 to the left, improve=3.2783080, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6509623, (0 missing)
## copd < 0.5 to the left, improve=0.5598170, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2946889, (0 missing)
##
## Node number 1931: 2247 observations, complexity param=0.0001605101
## predicted class=B2 expected loss=0.5959057 P(node) =0.008176767
## class counts: 795 908 383 148 13
## probabilities: 0.354 0.404 0.170 0.066 0.006
## left son=3862 (866 obs) right son=3863 (1381 obs)
## Primary splits:
## reimbursement2008 < 5335 to the right, improve=3.344298, (0 missing)
## copd < 0.5 to the left, improve=2.798571, (0 missing)
## age < 68.5 to the left, improve=2.255236, (0 missing)
## alzheimers < 0.5 to the left, improve=1.653597, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.448247, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.678, adj=0.165, (0 split)
## age < 34.5 to the left, agree=0.616, adj=0.005, (0 split)
##
## Node number 1932: 1987 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5546049 P(node) =0.007230634
## class counts: 662 885 320 111 9
## probabilities: 0.333 0.445 0.161 0.056 0.005
## left son=3864 (1964 obs) right son=3865 (23 obs)
## Primary splits:
## age < 98.5 to the left, improve=3.328502, (0 missing)
## heart.failure < 0.5 to the left, improve=3.141909, (0 missing)
## reimbursement2008 < 3085 to the left, improve=3.126917, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.906536, (0 missing)
## alzheimers < 0.5 to the left, improve=1.332632, (0 missing)
##
## Node number 1933: 941 observations
## predicted class=B2 expected loss=0.5494155 P(node) =0.003424271
## class counts: 242 424 186 79 10
## probabilities: 0.257 0.451 0.198 0.084 0.011
##
## Node number 1934: 36 observations
## predicted class=B1 expected loss=0.4444444 P(node) =0.0001310029
## class counts: 20 8 8 0 0
## probabilities: 0.556 0.222 0.222 0.000 0.000
##
## Node number 1935: 1428 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6239496 P(node) =0.00519645
## class counts: 455 537 282 146 8
## probabilities: 0.319 0.376 0.197 0.102 0.006
## left son=3870 (837 obs) right son=3871 (591 obs)
## Primary splits:
## age < 78.5 to the left, improve=2.474561, (0 missing)
## heart.failure < 0.5 to the left, improve=2.118405, (0 missing)
## stroke < 0.5 to the left, improve=1.930317, (0 missing)
## copd < 0.5 to the left, improve=1.447977, (0 missing)
## reimbursement2008 < 8670 to the right, improve=1.324274, (0 missing)
##
## Node number 1984: 101 observations
## predicted class=B2 expected loss=0.5247525 P(node) =0.000367536
## class counts: 33 48 15 4 1
## probabilities: 0.327 0.475 0.149 0.040 0.010
##
## Node number 1985: 471 observations, complexity param=5.811572e-05
## predicted class=B1 expected loss=0.5732484 P(node) =0.001713955
## class counts: 201 135 77 53 5
## probabilities: 0.427 0.287 0.163 0.113 0.011
## left son=3970 (346 obs) right son=3971 (125 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.506365, (0 missing)
## reimbursement2008 < 11515 to the left, improve=2.004779, (0 missing)
## heart.failure < 0.5 to the left, improve=1.922393, (0 missing)
## age < 72.5 to the right, improve=1.840715, (0 missing)
## bucket2008 < 2.5 to the left, improve=1.312872, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3600 to the right, agree=0.737, adj=0.008, (0 split)
##
## Node number 1986: 9 observations
## predicted class=B1 expected loss=0.2222222 P(node) =3.275073e-05
## class counts: 7 2 0 0 0
## probabilities: 0.778 0.222 0.000 0.000 0.000
##
## Node number 1987: 383 observations
## predicted class=B2 expected loss=0.5744125 P(node) =0.001393726
## class counts: 115 163 75 25 5
## probabilities: 0.300 0.426 0.196 0.065 0.013
##
## Node number 1990: 2424 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6575908 P(node) =0.008820864
## class counts: 663 830 547 335 49
## probabilities: 0.274 0.342 0.226 0.138 0.020
## left son=3980 (1234 obs) right son=3981 (1190 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.612677, (0 missing)
## age < 67.5 to the right, improve=3.329297, (0 missing)
## copd < 0.5 to the left, improve=3.109296, (0 missing)
## heart.failure < 0.5 to the left, improve=2.737555, (0 missing)
## reimbursement2008 < 9205 to the right, improve=2.610291, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.552, adj=0.087, (0 split)
## copd < 0.5 to the left, agree=0.540, adj=0.064, (0 split)
## age < 53.5 to the right, agree=0.526, adj=0.035, (0 split)
## reimbursement2008 < 12525 to the left, agree=0.522, adj=0.026, (0 split)
## heart.failure < 0.5 to the left, agree=0.520, adj=0.023, (0 split)
##
## Node number 1991: 1226 observations
## predicted class=B2 expected loss=0.5831974 P(node) =0.004461378
## class counts: 254 511 265 175 21
## probabilities: 0.207 0.417 0.216 0.143 0.017
##
## Node number 2018: 293 observations
## predicted class=B2 expected loss=0.4914676 P(node) =0.001066218
## class counts: 20 149 81 39 4
## probabilities: 0.068 0.509 0.276 0.133 0.014
##
## Node number 2019: 25 observations
## predicted class=B3 expected loss=0.4 P(node) =9.097426e-05
## class counts: 3 1 15 6 0
## probabilities: 0.120 0.040 0.600 0.240 0.000
##
## Node number 2030: 237 observations, complexity param=6.088314e-05
## predicted class=B4 expected loss=0.6329114 P(node) =0.000862436
## class counts: 13 76 49 87 12
## probabilities: 0.055 0.321 0.207 0.367 0.051
## left son=4060 (62 obs) right son=4061 (175 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=2.618202, (0 missing)
## reimbursement2008 < 90420 to the right, improve=2.488954, (0 missing)
## heart.failure < 0.5 to the left, improve=2.039633, (0 missing)
## age < 61.5 to the left, improve=1.881916, (0 missing)
## alzheimers < 0.5 to the left, improve=1.753135, (0 missing)
##
## Node number 2031: 14 observations
## predicted class=B4 expected loss=0.2142857 P(node) =5.094559e-05
## class counts: 0 1 1 11 1
## probabilities: 0.000 0.071 0.071 0.786 0.071
##
## Node number 2032: 72 observations
## predicted class=B1 expected loss=0.5694444 P(node) =0.0002620059
## class counts: 31 18 11 8 4
## probabilities: 0.431 0.250 0.153 0.111 0.056
##
## Node number 2033: 1245 observations
## predicted class=B2 expected loss=0.6714859 P(node) =0.004530518
## class counts: 238 409 223 305 70
## probabilities: 0.191 0.329 0.179 0.245 0.056
##
## Node number 2034: 191 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6753927 P(node) =0.0006950434
## class counts: 29 62 44 42 14
## probabilities: 0.152 0.325 0.230 0.220 0.073
## left son=4068 (172 obs) right son=4069 (19 obs)
## Primary splits:
## age < 64.5 to the right, improve=2.5420870, (0 missing)
## reimbursement2008 < 71460 to the left, improve=2.3514440, (0 missing)
## alzheimers < 0.5 to the left, improve=1.2597430, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9221356, (0 missing)
## ihd < 0.5 to the right, improve=0.7384918, (0 missing)
##
## Node number 2035: 981 observations, complexity param=9.962695e-05
## predicted class=B4 expected loss=0.6788991 P(node) =0.00356983
## class counts: 243 203 158 315 62
## probabilities: 0.248 0.207 0.161 0.321 0.063
## left son=4070 (468 obs) right son=4071 (513 obs)
## Primary splits:
## reimbursement2008 < 23175 to the left, improve=5.196818, (0 missing)
## alzheimers < 0.5 to the right, improve=3.174409, (0 missing)
## ihd < 0.5 to the left, improve=2.640760, (0 missing)
## heart.failure < 0.5 to the left, improve=1.689264, (0 missing)
## age < 97.5 to the right, improve=1.688586, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.777, adj=0.532, (0 split)
## heart.failure < 0.5 to the left, agree=0.534, adj=0.024, (0 split)
## age < 53.5 to the left, agree=0.531, adj=0.017, (0 split)
## stroke < 0.5 to the right, agree=0.525, adj=0.004, (0 split)
##
## Node number 3304: 265 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.490566 P(node) =0.0009643272
## class counts: 135 82 34 13 1
## probabilities: 0.509 0.309 0.128 0.049 0.004
## left son=6608 (251 obs) right son=6609 (14 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=3.807509, (0 missing)
## ihd < 0.5 to the left, improve=2.996787, (0 missing)
## cancer < 0.5 to the left, improve=2.863288, (0 missing)
## reimbursement2008 < 1815 to the left, improve=1.417998, (0 missing)
## depression < 0.5 to the left, improve=1.227469, (0 missing)
##
## Node number 3305: 290 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5931034 P(node) =0.001055301
## class counts: 118 111 42 16 3
## probabilities: 0.407 0.383 0.145 0.055 0.010
## left son=6610 (213 obs) right son=6611 (77 obs)
## Primary splits:
## age < 81.5 to the left, improve=1.9355560, (0 missing)
## reimbursement2008 < 2015 to the right, improve=1.1719950, (0 missing)
## ihd < 0.5 to the right, improve=0.8443893, (0 missing)
## stroke < 0.5 to the right, improve=0.5640543, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4757090, (0 missing)
##
## Node number 3470: 840 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5452381 P(node) =0.003056735
## class counts: 382 309 103 41 5
## probabilities: 0.455 0.368 0.123 0.049 0.006
## left son=6940 (71 obs) right son=6941 (769 obs)
## Primary splits:
## age < 54.5 to the left, improve=2.5793890, (0 missing)
## kidney < 0.5 to the left, improve=2.0319410, (0 missing)
## depression < 0.5 to the left, improve=1.3091120, (0 missing)
## reimbursement2008 < 2945 to the right, improve=1.1530320, (0 missing)
## copd < 0.5 to the left, improve=0.9762638, (0 missing)
##
## Node number 3471: 217 observations
## predicted class=B2 expected loss=0.5345622 P(node) =0.0007896566
## class counts: 76 101 34 6 0
## probabilities: 0.350 0.465 0.157 0.028 0.000
##
## Node number 3500: 1099 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.5814377 P(node) =0.003999229
## class counts: 460 388 168 76 7
## probabilities: 0.419 0.353 0.153 0.069 0.006
## left son=7000 (1074 obs) right son=7001 (25 obs)
## Primary splits:
## age < 95.5 to the left, improve=2.515661, (0 missing)
## copd < 0.5 to the left, improve=2.359857, (0 missing)
## cancer < 0.5 to the left, improve=1.641148, (0 missing)
## reimbursement2008 < 2575 to the right, improve=1.347245, (0 missing)
## stroke < 0.5 to the left, improve=1.145174, (0 missing)
##
## Node number 3501: 653 observations, complexity param=0.0001129105
## predicted class=B2 expected loss=0.6309342 P(node) =0.002376248
## class counts: 226 241 126 55 5
## probabilities: 0.346 0.369 0.193 0.084 0.008
## left son=7002 (303 obs) right son=7003 (350 obs)
## Primary splits:
## reimbursement2008 < 2655 to the left, improve=2.636734, (0 missing)
## cancer < 0.5 to the left, improve=1.461370, (0 missing)
## age < 55.5 to the right, improve=1.350106, (0 missing)
## alzheimers < 0.5 to the left, improve=1.189997, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.091246, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.548, adj=0.026, (0 split)
## copd < 0.5 to the right, agree=0.542, adj=0.013, (0 split)
## age < 47.5 to the left, agree=0.539, adj=0.007, (0 split)
##
## Node number 3508: 382 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.6230366 P(node) =0.001390087
## class counts: 142 144 74 18 4
## probabilities: 0.372 0.377 0.194 0.047 0.010
## left son=7016 (229 obs) right son=7017 (153 obs)
## Primary splits:
## depression < 0.5 to the left, improve=0.9679873, (0 missing)
## reimbursement2008 < 2275 to the left, improve=0.8225283, (0 missing)
## age < 75.5 to the right, improve=0.7274055, (0 missing)
## cancer < 0.5 to the right, improve=0.6895810, (0 missing)
## alzheimers < 0.5 to the right, improve=0.4333447, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.620, adj=0.052, (0 split)
## age < 50.5 to the right, agree=0.613, adj=0.033, (0 split)
##
## Node number 3509: 21 observations
## predicted class=B1 expected loss=0.5714286 P(node) =7.641838e-05
## class counts: 9 3 5 2 2
## probabilities: 0.429 0.143 0.238 0.095 0.095
##
## Node number 3532: 322 observations
## predicted class=B1 expected loss=0.5465839 P(node) =0.001171748
## class counts: 146 119 43 14 0
## probabilities: 0.453 0.370 0.134 0.043 0.000
##
## Node number 3533: 14 observations
## predicted class=B2 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 4 9 1 0 0
## probabilities: 0.286 0.643 0.071 0.000 0.000
##
## Node number 3544: 1008 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.5853175 P(node) =0.003668082
## class counts: 418 400 133 53 4
## probabilities: 0.415 0.397 0.132 0.053 0.004
## left son=7088 (275 obs) right son=7089 (733 obs)
## Primary splits:
## reimbursement2008 < 2535 to the right, improve=0.9732083, (0 missing)
## age < 39 to the left, improve=0.9699606, (0 missing)
## copd < 0.5 to the left, improve=0.8468269, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4615681, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4416739, (0 missing)
## Surrogate splits:
## age < 36.5 to the left, agree=0.728, adj=0.004, (0 split)
##
## Node number 3545: 49 observations
## predicted class=B2 expected loss=0.5918367 P(node) =0.0001783096
## class counts: 16 20 12 1 0
## probabilities: 0.327 0.408 0.245 0.020 0.000
##
## Node number 3604: 286 observations
## predicted class=B1 expected loss=0.4685315 P(node) =0.001040746
## class counts: 152 94 35 5 0
## probabilities: 0.531 0.329 0.122 0.017 0.000
##
## Node number 3605: 398 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5477387 P(node) =0.00144831
## class counts: 180 137 60 21 0
## probabilities: 0.452 0.344 0.151 0.053 0.000
## left son=7210 (340 obs) right son=7211 (58 obs)
## Primary splits:
## reimbursement2008 < 4700 to the right, improve=5.7797840, (0 missing)
## alzheimers < 0.5 to the left, improve=1.1372610, (0 missing)
## age < 34.5 to the left, improve=0.9964329, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.5848011, (0 missing)
## kidney < 0.5 to the right, improve=0.4151452, (0 missing)
##
## Node number 3624: 329 observations
## predicted class=B1 expected loss=0.4832827 P(node) =0.001197221
## class counts: 170 105 35 16 3
## probabilities: 0.517 0.319 0.106 0.049 0.009
##
## Node number 3625: 76 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5394737 P(node) =0.0002765618
## class counts: 32 35 7 2 0
## probabilities: 0.421 0.461 0.092 0.026 0.000
## left son=7250 (21 obs) right son=7251 (55 obs)
## Primary splits:
## reimbursement2008 < 6785 to the right, improve=3.2066300, (0 missing)
## bucket2008 < 2.5 to the right, improve=3.1159910, (0 missing)
## alzheimers < 0.5 to the right, improve=1.9967220, (0 missing)
## age < 85.5 to the right, improve=1.1176690, (0 missing)
## kidney < 0.5 to the right, improve=0.9258269, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.921, adj=0.714, (0 split)
## kidney < 0.5 to the right, agree=0.789, adj=0.238, (0 split)
##
## Node number 3626: 362 observations
## predicted class=B1 expected loss=0.5552486 P(node) =0.001317307
## class counts: 161 128 47 25 1
## probabilities: 0.445 0.354 0.130 0.069 0.003
##
## Node number 3627: 90 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5 P(node) =0.0003275073
## class counts: 29 45 11 5 0
## probabilities: 0.322 0.500 0.122 0.056 0.000
## left son=7254 (21 obs) right son=7255 (69 obs)
## Primary splits:
## age < 69.5 to the left, improve=3.1544510, (0 missing)
## alzheimers < 0.5 to the right, improve=3.1535260, (0 missing)
## kidney < 0.5 to the left, improve=1.7000000, (0 missing)
## reimbursement2008 < 3185 to the left, improve=1.5133190, (0 missing)
## copd < 0.5 to the right, improve=0.3083333, (0 missing)
##
## Node number 3654: 91 observations
## predicted class=B1 expected loss=0.5054945 P(node) =0.0003311463
## class counts: 45 31 9 6 0
## probabilities: 0.495 0.341 0.099 0.066 0.000
##
## Node number 3655: 266 observations
## predicted class=B2 expected loss=0.556391 P(node) =0.0009679661
## class counts: 95 118 39 13 1
## probabilities: 0.357 0.444 0.147 0.049 0.004
##
## Node number 3696: 105 observations
## predicted class=B1 expected loss=0.5238095 P(node) =0.0003820919
## class counts: 50 35 16 3 1
## probabilities: 0.476 0.333 0.152 0.029 0.010
##
## Node number 3697: 27 observations
## predicted class=B3 expected loss=0.6296296 P(node) =9.82522e-05
## class counts: 4 7 10 6 0
## probabilities: 0.148 0.259 0.370 0.222 0.000
##
## Node number 3850: 137 observations
## predicted class=B1 expected loss=0.4963504 P(node) =0.000498539
## class counts: 69 41 17 9 1
## probabilities: 0.504 0.299 0.124 0.066 0.007
##
## Node number 3851: 372 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5913978 P(node) =0.001353697
## class counts: 152 144 50 24 2
## probabilities: 0.409 0.387 0.134 0.065 0.005
## left son=7702 (330 obs) right son=7703 (42 obs)
## Primary splits:
## reimbursement2008 < 4055 to the right, improve=3.5107950, (0 missing)
## age < 96 to the left, improve=2.0766610, (0 missing)
## depression < 0.5 to the left, improve=1.3295540, (0 missing)
## copd < 0.5 to the left, improve=1.2612770, (0 missing)
## heart.failure < 0.5 to the left, improve=0.1950857, (0 missing)
##
## Node number 3852: 211 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.563981 P(node) =0.0007678228
## class counts: 92 89 23 7 0
## probabilities: 0.436 0.422 0.109 0.033 0.000
## left son=7704 (142 obs) right son=7705 (69 obs)
## Primary splits:
## reimbursement2008 < 4075 to the left, improve=3.4884480, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9268848, (0 missing)
## heart.failure < 0.5 to the left, improve=0.8762896, (0 missing)
## age < 95 to the right, improve=0.6396384, (0 missing)
## stroke < 0.5 to the left, improve=0.3979516, (0 missing)
## Surrogate splits:
## age < 96.5 to the left, agree=0.687, adj=0.043, (0 split)
##
## Node number 3853: 128 observations
## predicted class=B1 expected loss=0.546875 P(node) =0.0004657882
## class counts: 58 38 22 9 1
## probabilities: 0.453 0.297 0.172 0.070 0.008
##
## Node number 3854: 181 observations
## predicted class=B2 expected loss=0.4861878 P(node) =0.0006586537
## class counts: 56 93 23 7 2
## probabilities: 0.309 0.514 0.127 0.039 0.011
##
## Node number 3855: 58 observations
## predicted class=B3 expected loss=0.5862069 P(node) =0.0002110603
## class counts: 14 16 24 4 0
## probabilities: 0.241 0.276 0.414 0.069 0.000
##
## Node number 3858: 116 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.5517241 P(node) =0.0004221206
## class counts: 41 52 14 7 2
## probabilities: 0.353 0.448 0.121 0.060 0.017
## left son=7716 (63 obs) right son=7717 (53 obs)
## Primary splits:
## age < 74.5 to the right, improve=2.8010500, (0 missing)
## reimbursement2008 < 17265 to the left, improve=2.4722090, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.9776500, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5892240, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5845524, (0 missing)
## Surrogate splits:
## reimbursement2008 < 15590 to the right, agree=0.603, adj=0.132, (0 split)
## heart.failure < 0.5 to the right, agree=0.578, adj=0.075, (0 split)
## alzheimers < 0.5 to the right, agree=0.560, adj=0.038, (0 split)
## osteoporosis < 0.5 to the left, agree=0.552, adj=0.019, (0 split)
## bucket2008 < 3.5 to the right, agree=0.552, adj=0.019, (0 split)
##
## Node number 3859: 449 observations
## predicted class=B1 expected loss=0.6191537 P(node) =0.001633898
## class counts: 171 127 110 34 7
## probabilities: 0.381 0.283 0.245 0.076 0.016
##
## Node number 3860: 343 observations
## predicted class=B1 expected loss=0.5014577 P(node) =0.001248167
## class counts: 171 126 33 11 2
## probabilities: 0.499 0.367 0.096 0.032 0.006
##
## Node number 3861: 1610 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5925466 P(node) =0.005858742
## class counts: 653 656 218 77 6
## probabilities: 0.406 0.407 0.135 0.048 0.004
## left son=7722 (43 obs) right son=7723 (1567 obs)
## Primary splits:
## age < 42.5 to the left, improve=2.9787580, (0 missing)
## reimbursement2008 < 3475 to the right, improve=1.6291410, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4315089, (0 missing)
## copd < 0.5 to the left, improve=0.2703192, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2492479, (0 missing)
##
## Node number 3862: 866 observations, complexity param=0.0001605101
## predicted class=B1 expected loss=0.6120092 P(node) =0.003151348
## class counts: 336 322 139 64 5
## probabilities: 0.388 0.372 0.161 0.074 0.006
## left son=7724 (129 obs) right son=7725 (737 obs)
## Primary splits:
## reimbursement2008 < 8115 to the right, improve=2.0175360, (0 missing)
## age < 89.5 to the left, improve=1.8274460, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.6689370, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.3462440, (0 missing)
## stroke < 0.5 to the left, improve=0.5970317, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.984, adj=0.891, (0 split)
##
## Node number 3863: 1381 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.5756698 P(node) =0.005025418
## class counts: 459 586 244 84 8
## probabilities: 0.332 0.424 0.177 0.061 0.006
## left son=7726 (997 obs) right son=7727 (384 obs)
## Primary splits:
## copd < 0.5 to the left, improve=3.5627620, (0 missing)
## age < 37.5 to the right, improve=2.2016010, (0 missing)
## reimbursement2008 < 5195 to the left, improve=1.9417980, (0 missing)
## alzheimers < 0.5 to the left, improve=1.9283600, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3841814, (0 missing)
## Surrogate splits:
## age < 34 to the right, agree=0.723, adj=0.005, (0 split)
## reimbursement2008 < 5325 to the left, agree=0.723, adj=0.005, (0 split)
##
## Node number 3864: 1964 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.552444 P(node) =0.007146938
## class counts: 656 879 309 111 9
## probabilities: 0.334 0.448 0.157 0.057 0.005
## left son=7728 (22 obs) right son=7729 (1942 obs)
## Primary splits:
## reimbursement2008 < 3085 to the left, improve=3.418849, (0 missing)
## heart.failure < 0.5 to the left, improve=3.308540, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.919418, (0 missing)
## age < 66.5 to the right, improve=1.961336, (0 missing)
## alzheimers < 0.5 to the left, improve=1.448295, (0 missing)
##
## Node number 3865: 23 observations
## predicted class=B3 expected loss=0.5217391 P(node) =8.369632e-05
## class counts: 6 6 11 0 0
## probabilities: 0.261 0.261 0.478 0.000 0.000
##
## Node number 3870: 837 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6356033 P(node) =0.003045818
## class counts: 292 305 155 82 3
## probabilities: 0.349 0.364 0.185 0.098 0.004
## left son=7740 (639 obs) right son=7741 (198 obs)
## Primary splits:
## reimbursement2008 < 21320 to the left, improve=1.8992410, (0 missing)
## alzheimers < 0.5 to the left, improve=1.6748470, (0 missing)
## age < 49.5 to the left, improve=1.4981360, (0 missing)
## bucket2008 < 3.5 to the left, improve=1.4460210, (0 missing)
## stroke < 0.5 to the right, improve=0.7571134, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.955, adj=0.808, (0 split)
##
## Node number 3871: 591 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.607445 P(node) =0.002150632
## class counts: 163 232 127 64 5
## probabilities: 0.276 0.393 0.215 0.108 0.008
## left son=7742 (122 obs) right son=7743 (469 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.4607800, (0 missing)
## reimbursement2008 < 8775 to the left, improve=1.9199300, (0 missing)
## stroke < 0.5 to the left, improve=1.2822570, (0 missing)
## copd < 0.5 to the left, improve=1.1325150, (0 missing)
## age < 80.5 to the right, improve=0.6677683, (0 missing)
##
## Node number 3970: 346 observations
## predicted class=B1 expected loss=0.5375723 P(node) =0.001259084
## class counts: 160 92 52 37 5
## probabilities: 0.462 0.266 0.150 0.107 0.014
##
## Node number 3971: 125 observations, complexity param=5.811572e-05
## predicted class=B2 expected loss=0.656 P(node) =0.0004548713
## class counts: 41 43 25 16 0
## probabilities: 0.328 0.344 0.200 0.128 0.000
## left son=7942 (106 obs) right son=7943 (19 obs)
## Primary splits:
## age < 62 to the right, improve=3.3415930, (0 missing)
## reimbursement2008 < 11475 to the left, improve=2.2730020, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7920000, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.7319750, (0 missing)
## stroke < 0.5 to the right, improve=0.5402967, (0 missing)
##
## Node number 3980: 1234 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6612642 P(node) =0.00449049
## class counts: 374 418 252 160 30
## probabilities: 0.303 0.339 0.204 0.130 0.024
## left son=7960 (349 obs) right son=7961 (885 obs)
## Primary splits:
## reimbursement2008 < 12135 to the right, improve=3.745241, (0 missing)
## age < 67.5 to the left, improve=3.421516, (0 missing)
## heart.failure < 0.5 to the left, improve=1.338981, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.254047, (0 missing)
## copd < 0.5 to the right, improve=1.093433, (0 missing)
##
## Node number 3981: 1190 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6537815 P(node) =0.004330375
## class counts: 289 412 295 175 19
## probabilities: 0.243 0.346 0.248 0.147 0.016
## left son=7962 (547 obs) right son=7963 (643 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.655589, (0 missing)
## age < 82.5 to the right, improve=1.853724, (0 missing)
## stroke < 0.5 to the right, improve=1.583996, (0 missing)
## reimbursement2008 < 6355 to the right, improve=1.325969, (0 missing)
## heart.failure < 0.5 to the left, improve=1.085887, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the left, agree=0.562, adj=0.048, (0 split)
## reimbursement2008 < 7315 to the left, agree=0.555, adj=0.031, (0 split)
## alzheimers < 0.5 to the left, agree=0.543, adj=0.005, (0 split)
## age < 28.5 to the left, agree=0.542, adj=0.004, (0 split)
##
## Node number 4060: 62 observations
## predicted class=B2 expected loss=0.5806452 P(node) =0.0002256162
## class counts: 3 26 17 15 1
## probabilities: 0.048 0.419 0.274 0.242 0.016
##
## Node number 4061: 175 observations
## predicted class=B4 expected loss=0.5885714 P(node) =0.0006368198
## class counts: 10 50 32 72 11
## probabilities: 0.057 0.286 0.183 0.411 0.063
##
## Node number 4068: 172 observations
## predicted class=B2 expected loss=0.6511628 P(node) =0.0006259029
## class counts: 27 60 35 39 11
## probabilities: 0.157 0.349 0.203 0.227 0.064
##
## Node number 4069: 19 observations
## predicted class=B3 expected loss=0.5263158 P(node) =6.914044e-05
## class counts: 2 2 9 3 3
## probabilities: 0.105 0.105 0.474 0.158 0.158
##
## Node number 4070: 468 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7200855 P(node) =0.001703038
## class counts: 131 112 77 122 26
## probabilities: 0.280 0.239 0.165 0.261 0.056
## left son=8140 (457 obs) right son=8141 (11 obs)
## Primary splits:
## age < 93.5 to the left, improve=1.955850, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.902288, (0 missing)
## reimbursement2008 < 19700 to the right, improve=1.873153, (0 missing)
## ihd < 0.5 to the right, improve=1.868954, (0 missing)
## alzheimers < 0.5 to the right, improve=1.716285, (0 missing)
##
## Node number 4071: 513 observations
## predicted class=B4 expected loss=0.6237817 P(node) =0.001866792
## class counts: 112 91 81 193 36
## probabilities: 0.218 0.177 0.158 0.376 0.070
##
## Node number 6608: 251 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.4701195 P(node) =0.0009133816
## class counts: 133 73 33 11 1
## probabilities: 0.530 0.291 0.131 0.044 0.004
## left son=13216 (235 obs) right son=13217 (16 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.6866090, (0 missing)
## ihd < 0.5 to the left, improve=2.4584220, (0 missing)
## reimbursement2008 < 1815 to the left, improve=1.2636760, (0 missing)
## age < 60.5 to the right, improve=1.0616730, (0 missing)
## depression < 0.5 to the left, improve=0.7871996, (0 missing)
##
## Node number 6609: 14 observations
## predicted class=B2 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 2 9 1 2 0
## probabilities: 0.143 0.643 0.071 0.143 0.000
##
## Node number 6610: 213 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5633803 P(node) =0.0007751007
## class counts: 93 74 30 15 1
## probabilities: 0.437 0.347 0.141 0.070 0.005
## left son=13220 (201 obs) right son=13221 (12 obs)
## Primary splits:
## age < 44.5 to the right, improve=1.7572700, (0 missing)
## reimbursement2008 < 2005 to the right, improve=1.0657280, (0 missing)
## cancer < 0.5 to the right, improve=0.3892571, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3640621, (0 missing)
## depression < 0.5 to the right, improve=0.3467310, (0 missing)
##
## Node number 6611: 77 observations
## predicted class=B2 expected loss=0.5194805 P(node) =0.0002802007
## class counts: 25 37 12 1 2
## probabilities: 0.325 0.481 0.156 0.013 0.026
##
## Node number 6940: 71 observations
## predicted class=B1 expected loss=0.4366197 P(node) =0.0002583669
## class counts: 40 16 11 3 1
## probabilities: 0.563 0.225 0.155 0.042 0.014
##
## Node number 6941: 769 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5552666 P(node) =0.002798368
## class counts: 342 293 92 38 4
## probabilities: 0.445 0.381 0.120 0.049 0.005
## left son=13882 (472 obs) right son=13883 (297 obs)
## Primary splits:
## age < 70.5 to the right, improve=2.5248320, (0 missing)
## depression < 0.5 to the left, improve=1.8557100, (0 missing)
## kidney < 0.5 to the left, improve=1.7236880, (0 missing)
## reimbursement2008 < 2665 to the right, improve=1.1252400, (0 missing)
## copd < 0.5 to the left, improve=0.9387137, (0 missing)
##
## Node number 7000: 1074 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5772812 P(node) =0.003908254
## class counts: 454 373 166 74 7
## probabilities: 0.423 0.347 0.155 0.069 0.007
## left son=14000 (808 obs) right son=14001 (266 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.7468870, (0 missing)
## cancer < 0.5 to the left, improve=1.8315690, (0 missing)
## age < 78.5 to the left, improve=1.6074350, (0 missing)
## reimbursement2008 < 2575 to the right, improve=1.2651380, (0 missing)
## stroke < 0.5 to the left, improve=0.9951466, (0 missing)
##
## Node number 7001: 25 observations
## predicted class=B2 expected loss=0.4 P(node) =9.097426e-05
## class counts: 6 15 2 2 0
## probabilities: 0.240 0.600 0.080 0.080 0.000
##
## Node number 7002: 303 observations
## predicted class=B1 expected loss=0.6039604 P(node) =0.001102608
## class counts: 120 99 62 21 1
## probabilities: 0.396 0.327 0.205 0.069 0.003
##
## Node number 7003: 350 observations
## predicted class=B2 expected loss=0.5942857 P(node) =0.00127364
## class counts: 106 142 64 34 4
## probabilities: 0.303 0.406 0.183 0.097 0.011
##
## Node number 7016: 229 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5938865 P(node) =0.0008333242
## class counts: 93 82 44 8 2
## probabilities: 0.406 0.358 0.192 0.035 0.009
## left son=14032 (15 obs) right son=14033 (214 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=1.9222650, (0 missing)
## reimbursement2008 < 2515 to the left, improve=1.4164780, (0 missing)
## age < 94 to the left, improve=1.2547820, (0 missing)
## alzheimers < 0.5 to the right, improve=0.6631197, (0 missing)
## copd < 0.5 to the right, improve=0.2469242, (0 missing)
##
## Node number 7017: 153 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.5947712 P(node) =0.0005567625
## class counts: 49 62 30 10 2
## probabilities: 0.320 0.405 0.196 0.065 0.013
## left son=14034 (14 obs) right son=14035 (139 obs)
## Primary splits:
## reimbursement2008 < 2545 to the right, improve=2.7113570, (0 missing)
## age < 45 to the left, improve=1.6972360, (0 missing)
## cancer < 0.5 to the left, improve=0.6348039, (0 missing)
## copd < 0.5 to the right, improve=0.3887797, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2839287, (0 missing)
##
## Node number 7088: 275 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.56 P(node) =0.001000717
## class counts: 108 121 34 11 1
## probabilities: 0.393 0.440 0.124 0.040 0.004
## left son=14176 (44 obs) right son=14177 (231 obs)
## Primary splits:
## age < 63.5 to the left, improve=2.2068400, (0 missing)
## reimbursement2008 < 2555 to the right, improve=1.7374730, (0 missing)
## alzheimers < 0.5 to the right, improve=1.5968660, (0 missing)
## copd < 0.5 to the left, improve=0.9046397, (0 missing)
## heart.failure < 0.5 to the right, improve=0.5279104, (0 missing)
##
## Node number 7089: 733 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5770805 P(node) =0.002667365
## class counts: 310 279 99 42 3
## probabilities: 0.423 0.381 0.135 0.057 0.004
## left son=14178 (10 obs) right son=14179 (723 obs)
## Primary splits:
## age < 97.5 to the right, improve=1.1550490, (0 missing)
## heart.failure < 0.5 to the left, improve=1.1107930, (0 missing)
## reimbursement2008 < 2495 to the right, improve=0.7495829, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7242328, (0 missing)
## depression < 0.5 to the left, improve=0.5684301, (0 missing)
##
## Node number 7210: 340 observations
## predicted class=B1 expected loss=0.5088235 P(node) =0.00123725
## class counts: 167 107 48 18 0
## probabilities: 0.491 0.315 0.141 0.053 0.000
##
## Node number 7211: 58 observations
## predicted class=B2 expected loss=0.4827586 P(node) =0.0002110603
## class counts: 13 30 12 3 0
## probabilities: 0.224 0.517 0.207 0.052 0.000
##
## Node number 7250: 21 observations
## predicted class=B1 expected loss=0.3333333 P(node) =7.641838e-05
## class counts: 14 5 2 0 0
## probabilities: 0.667 0.238 0.095 0.000 0.000
##
## Node number 7251: 55 observations
## predicted class=B2 expected loss=0.4545455 P(node) =0.0002001434
## class counts: 18 30 5 2 0
## probabilities: 0.327 0.545 0.091 0.036 0.000
##
## Node number 7254: 21 observations
## predicted class=B1 expected loss=0.4285714 P(node) =7.641838e-05
## class counts: 12 6 1 2 0
## probabilities: 0.571 0.286 0.048 0.095 0.000
##
## Node number 7255: 69 observations
## predicted class=B2 expected loss=0.4347826 P(node) =0.000251089
## class counts: 17 39 10 3 0
## probabilities: 0.246 0.565 0.145 0.043 0.000
##
## Node number 7702: 330 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.569697 P(node) =0.00120086
## class counts: 142 119 44 23 2
## probabilities: 0.430 0.361 0.133 0.070 0.006
## left son=15404 (309 obs) right son=15405 (21 obs)
## Primary splits:
## reimbursement2008 < 4185 to the right, improve=2.2681710, (0 missing)
## age < 96 to the left, improve=2.1333520, (0 missing)
## depression < 0.5 to the left, improve=0.7533962, (0 missing)
## copd < 0.5 to the left, improve=0.6700147, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.3769465, (0 missing)
##
## Node number 7703: 42 observations
## predicted class=B2 expected loss=0.4047619 P(node) =0.0001528368
## class counts: 10 25 6 1 0
## probabilities: 0.238 0.595 0.143 0.024 0.000
##
## Node number 7704: 142 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0005167338
## class counts: 71 51 15 5 0
## probabilities: 0.500 0.359 0.106 0.035 0.000
##
## Node number 7705: 69 observations
## predicted class=B2 expected loss=0.4492754 P(node) =0.000251089
## class counts: 21 38 8 2 0
## probabilities: 0.304 0.551 0.116 0.029 0.000
##
## Node number 7716: 63 observations
## predicted class=B1 expected loss=0.5714286 P(node) =0.0002292551
## class counts: 27 21 10 4 1
## probabilities: 0.429 0.333 0.159 0.063 0.016
##
## Node number 7717: 53 observations
## predicted class=B2 expected loss=0.4150943 P(node) =0.0001928654
## class counts: 14 31 4 3 1
## probabilities: 0.264 0.585 0.075 0.057 0.019
##
## Node number 7722: 43 observations
## predicted class=B1 expected loss=0.3953488 P(node) =0.0001564757
## class counts: 26 11 3 3 0
## probabilities: 0.605 0.256 0.070 0.070 0.000
##
## Node number 7723: 1567 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5883854 P(node) =0.005702267
## class counts: 627 645 215 74 6
## probabilities: 0.400 0.412 0.137 0.047 0.004
## left son=15446 (1527 obs) right son=15447 (40 obs)
## Primary splits:
## age < 50.5 to the right, improve=1.7032880, (0 missing)
## reimbursement2008 < 3475 to the right, improve=1.5459000, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4552758, (0 missing)
## stroke < 0.5 to the left, improve=0.2471234, (0 missing)
## copd < 0.5 to the left, improve=0.2014160, (0 missing)
##
## Node number 7724: 129 observations
## predicted class=B2 expected loss=0.5271318 P(node) =0.0004694272
## class counts: 46 61 16 6 0
## probabilities: 0.357 0.473 0.124 0.047 0.000
##
## Node number 7725: 737 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.6065129 P(node) =0.002681921
## class counts: 290 261 123 58 5
## probabilities: 0.393 0.354 0.167 0.079 0.007
## left son=15450 (703 obs) right son=15451 (34 obs)
## Primary splits:
## age < 94.5 to the left, improve=1.9050170, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.5567680, (0 missing)
## reimbursement2008 < 6575 to the right, improve=1.5078350, (0 missing)
## copd < 0.5 to the left, improve=0.5423379, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5213862, (0 missing)
##
## Node number 7726: 997 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.5927783 P(node) =0.003628054
## class counts: 357 406 171 57 6
## probabilities: 0.358 0.407 0.172 0.057 0.006
## left son=15452 (297 obs) right son=15453 (700 obs)
## Primary splits:
## age < 69.5 to the left, improve=2.4458440, (0 missing)
## alzheimers < 0.5 to the left, improve=2.2624190, (0 missing)
## reimbursement2008 < 4135 to the right, improve=1.8635870, (0 missing)
## stroke < 0.5 to the right, improve=0.3191114, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3114115, (0 missing)
## Surrogate splits:
## reimbursement2008 < 5325 to the right, agree=0.703, adj=0.003, (0 split)
##
## Node number 7727: 384 observations
## predicted class=B2 expected loss=0.53125 P(node) =0.001397365
## class counts: 102 180 73 27 2
## probabilities: 0.266 0.469 0.190 0.070 0.005
##
## Node number 7728: 22 observations
## predicted class=B1 expected loss=0.3636364 P(node) =8.005735e-05
## class counts: 14 5 1 2 0
## probabilities: 0.636 0.227 0.045 0.091 0.000
##
## Node number 7729: 1942 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5499485 P(node) =0.007066881
## class counts: 642 874 308 109 9
## probabilities: 0.331 0.450 0.159 0.056 0.005
## left son=15458 (889 obs) right son=15459 (1053 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.215424, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.693064, (0 missing)
## reimbursement2008 < 8025 to the left, improve=2.054172, (0 missing)
## age < 66.5 to the right, improve=1.953237, (0 missing)
## alzheimers < 0.5 to the left, improve=1.238773, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3815 to the left, agree=0.545, adj=0.007, (0 split)
## age < 31.5 to the left, agree=0.544, adj=0.003, (0 split)
##
## Node number 7740: 639 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6353678 P(node) =0.002325302
## class counts: 233 221 124 59 2
## probabilities: 0.365 0.346 0.194 0.092 0.003
## left son=15480 (83 obs) right son=15481 (556 obs)
## Primary splits:
## age < 49.5 to the left, improve=1.7453130, (0 missing)
## stroke < 0.5 to the left, improve=1.0790130, (0 missing)
## reimbursement2008 < 10445 to the right, improve=1.0644190, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9731198, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.6917546, (0 missing)
##
## Node number 7741: 198 observations
## predicted class=B2 expected loss=0.5757576 P(node) =0.0007205162
## class counts: 59 84 31 23 1
## probabilities: 0.298 0.424 0.157 0.116 0.005
##
## Node number 7742: 122 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5983607 P(node) =0.0004439544
## class counts: 49 39 22 12 0
## probabilities: 0.402 0.320 0.180 0.098 0.000
## left son=15484 (40 obs) right son=15485 (82 obs)
## Primary splits:
## reimbursement2008 < 11560 to the left, improve=2.7817470, (0 missing)
## age < 80.5 to the right, improve=1.9103730, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4632510, (0 missing)
## copd < 0.5 to the left, improve=1.0641520, (0 missing)
## stroke < 0.5 to the left, improve=0.9890302, (0 missing)
##
## Node number 7743: 469 observations
## predicted class=B2 expected loss=0.5884861 P(node) =0.001706677
## class counts: 114 193 105 52 5
## probabilities: 0.243 0.412 0.224 0.111 0.011
##
## Node number 7942: 106 observations, complexity param=5.811572e-05
## predicted class=B1 expected loss=0.6320755 P(node) =0.0003857309
## class counts: 39 39 16 12 0
## probabilities: 0.368 0.368 0.151 0.113 0.000
## left son=15884 (93 obs) right son=15885 (13 obs)
## Primary splits:
## age < 67.5 to the right, improve=2.3273090, (0 missing)
## reimbursement2008 < 11575 to the left, improve=1.8244140, (0 missing)
## stroke < 0.5 to the right, improve=0.5034792, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4237564, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3937905, (0 missing)
##
## Node number 7943: 19 observations
## predicted class=B3 expected loss=0.5263158 P(node) =6.914044e-05
## class counts: 2 4 9 4 0
## probabilities: 0.105 0.211 0.474 0.211 0.000
##
## Node number 7960: 349 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6475645 P(node) =0.001270001
## class counts: 123 103 55 57 11
## probabilities: 0.352 0.295 0.158 0.163 0.032
## left son=15920 (331 obs) right son=15921 (18 obs)
## Primary splits:
## age < 54 to the right, improve=2.0196730, (0 missing)
## alzheimers < 0.5 to the left, improve=1.9958000, (0 missing)
## reimbursement2008 < 15235 to the left, improve=1.7314800, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9381122, (0 missing)
## copd < 0.5 to the right, improve=0.4818854, (0 missing)
##
## Node number 7961: 885 observations
## predicted class=B2 expected loss=0.6440678 P(node) =0.003220489
## class counts: 251 315 197 103 19
## probabilities: 0.284 0.356 0.223 0.116 0.021
##
## Node number 7962: 547 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6709324 P(node) =0.001990517
## class counts: 154 180 136 65 12
## probabilities: 0.282 0.329 0.249 0.119 0.022
## left son=15924 (310 obs) right son=15925 (237 obs)
## Primary splits:
## reimbursement2008 < 9205 to the right, improve=2.7787810, (0 missing)
## age < 68.5 to the right, improve=2.5123800, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.9624740, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6055315, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5192530, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.835, adj=0.620, (0 split)
## age < 44.5 to the right, agree=0.581, adj=0.034, (0 split)
##
## Node number 7963: 643 observations
## predicted class=B2 expected loss=0.6391913 P(node) =0.002339858
## class counts: 135 232 159 110 7
## probabilities: 0.210 0.361 0.247 0.171 0.011
##
## Node number 8140: 457 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7133479 P(node) =0.00166301
## class counts: 131 107 73 120 26
## probabilities: 0.287 0.234 0.160 0.263 0.057
## left son=16280 (398 obs) right son=16281 (59 obs)
## Primary splits:
## age < 86.5 to the left, improve=2.218362, (0 missing)
## ihd < 0.5 to the right, improve=2.044330, (0 missing)
## reimbursement2008 < 19430 to the right, improve=1.853412, (0 missing)
## alzheimers < 0.5 to the right, improve=1.735004, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.352826, (0 missing)
##
## Node number 8141: 11 observations
## predicted class=B2 expected loss=0.5454545 P(node) =4.002868e-05
## class counts: 0 5 4 2 0
## probabilities: 0.000 0.455 0.364 0.182 0.000
##
## Node number 13216: 235 observations
## predicted class=B1 expected loss=0.4510638 P(node) =0.0008551581
## class counts: 129 64 30 11 1
## probabilities: 0.549 0.272 0.128 0.047 0.004
##
## Node number 13217: 16 observations
## predicted class=B2 expected loss=0.4375 P(node) =5.822353e-05
## class counts: 4 9 3 0 0
## probabilities: 0.250 0.562 0.188 0.000 0.000
##
## Node number 13220: 201 observations
## predicted class=B1 expected loss=0.5472637 P(node) =0.0007314331
## class counts: 91 67 29 14 0
## probabilities: 0.453 0.333 0.144 0.070 0.000
##
## Node number 13221: 12 observations
## predicted class=B2 expected loss=0.4166667 P(node) =4.366765e-05
## class counts: 2 7 1 1 1
## probabilities: 0.167 0.583 0.083 0.083 0.083
##
## Node number 13882: 472 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5275424 P(node) =0.001717594
## class counts: 223 163 57 25 4
## probabilities: 0.472 0.345 0.121 0.053 0.008
## left son=27764 (343 obs) right son=27765 (129 obs)
## Primary splits:
## age < 73.5 to the right, improve=4.630612, (0 missing)
## reimbursement2008 < 2805 to the right, improve=1.597068, (0 missing)
## depression < 0.5 to the left, improve=1.459900, (0 missing)
## kidney < 0.5 to the left, improve=1.335760, (0 missing)
## stroke < 0.5 to the left, improve=1.130037, (0 missing)
##
## Node number 13883: 297 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.5622896 P(node) =0.001080774
## class counts: 119 130 35 13 0
## probabilities: 0.401 0.438 0.118 0.044 0.000
## left son=27766 (218 obs) right son=27767 (79 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=1.3951400, (0 missing)
## reimbursement2008 < 2945 to the right, improve=1.0350230, (0 missing)
## depression < 0.5 to the left, improve=0.9259259, (0 missing)
## kidney < 0.5 to the left, improve=0.7583938, (0 missing)
## copd < 0.5 to the left, improve=0.3569379, (0 missing)
##
## Node number 14000: 808 observations
## predicted class=B1 expected loss=0.5569307 P(node) =0.002940288
## class counts: 358 273 111 61 5
## probabilities: 0.443 0.338 0.137 0.075 0.006
##
## Node number 14001: 266 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.6240602 P(node) =0.0009679661
## class counts: 96 100 55 13 2
## probabilities: 0.361 0.376 0.207 0.049 0.008
## left son=28002 (192 obs) right son=28003 (74 obs)
## Primary splits:
## reimbursement2008 < 2540 to the right, improve=2.9691060, (0 missing)
## age < 78.5 to the left, improve=2.6852920, (0 missing)
## cancer < 0.5 to the right, improve=2.3754980, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7018574, (0 missing)
## stroke < 0.5 to the left, improve=0.6157537, (0 missing)
## Surrogate splits:
## age < 50.5 to the right, agree=0.737, adj=0.054, (0 split)
##
## Node number 14032: 15 observations
## predicted class=B1 expected loss=0.3333333 P(node) =5.458456e-05
## class counts: 10 2 3 0 0
## probabilities: 0.667 0.133 0.200 0.000 0.000
##
## Node number 14033: 214 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6121495 P(node) =0.0007787397
## class counts: 83 80 41 8 2
## probabilities: 0.388 0.374 0.192 0.037 0.009
## left son=28066 (169 obs) right son=28067 (45 obs)
## Primary splits:
## reimbursement2008 < 2515 to the left, improve=1.6030020, (0 missing)
## age < 52.5 to the right, improve=0.9765448, (0 missing)
## alzheimers < 0.5 to the right, improve=0.7668533, (0 missing)
## copd < 0.5 to the right, improve=0.3681910, (0 missing)
## ihd < 0.5 to the right, improve=0.1207875, (0 missing)
##
## Node number 14034: 14 observations
## predicted class=B1 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 9 2 2 1 0
## probabilities: 0.643 0.143 0.143 0.071 0.000
##
## Node number 14035: 139 observations
## predicted class=B2 expected loss=0.5683453 P(node) =0.0005058169
## class counts: 40 60 28 9 2
## probabilities: 0.288 0.432 0.201 0.065 0.014
##
## Node number 14176: 44 observations
## predicted class=B2 expected loss=0.3863636 P(node) =0.0001601147
## class counts: 14 27 2 1 0
## probabilities: 0.318 0.614 0.045 0.023 0.000
##
## Node number 14177: 231 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.5930736 P(node) =0.0008406022
## class counts: 94 94 32 10 1
## probabilities: 0.407 0.407 0.139 0.043 0.004
## left son=28354 (169 obs) right son=28355 (62 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.4583990, (0 missing)
## reimbursement2008 < 2555 to the right, improve=1.0376560, (0 missing)
## age < 84.5 to the left, improve=1.0243680, (0 missing)
## copd < 0.5 to the left, improve=0.7240171, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4410819, (0 missing)
## Surrogate splits:
## age < 87.5 to the left, agree=0.745, adj=0.048, (0 split)
##
## Node number 14178: 10 observations
## predicted class=B1 expected loss=0.3 P(node) =3.63897e-05
## class counts: 7 2 1 0 0
## probabilities: 0.700 0.200 0.100 0.000 0.000
##
## Node number 14179: 723 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5809129 P(node) =0.002630976
## class counts: 303 277 98 42 3
## probabilities: 0.419 0.383 0.136 0.058 0.004
## left son=28358 (689 obs) right son=28359 (34 obs)
## Primary splits:
## age < 90.5 to the left, improve=1.6650270, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5078050, (0 missing)
## reimbursement2008 < 2495 to the right, improve=0.8133392, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6699213, (0 missing)
## depression < 0.5 to the left, improve=0.5296598, (0 missing)
##
## Node number 15404: 309 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5728155 P(node) =0.001124442
## class counts: 132 117 38 20 2
## probabilities: 0.427 0.379 0.123 0.065 0.006
## left son=30808 (253 obs) right son=30809 (56 obs)
## Primary splits:
## reimbursement2008 < 4635 to the right, improve=2.0908250, (0 missing)
## age < 73.5 to the right, improve=1.8355900, (0 missing)
## depression < 0.5 to the left, improve=0.6554201, (0 missing)
## copd < 0.5 to the left, improve=0.3380891, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.2757170, (0 missing)
##
## Node number 15405: 21 observations
## predicted class=B1 expected loss=0.5238095 P(node) =7.641838e-05
## class counts: 10 2 6 3 0
## probabilities: 0.476 0.095 0.286 0.143 0.000
##
## Node number 15446: 1527 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5854617 P(node) =0.005556708
## class counts: 613 633 203 72 6
## probabilities: 0.401 0.415 0.133 0.047 0.004
## left son=30892 (1478 obs) right son=30893 (49 obs)
## Primary splits:
## reimbursement2008 < 3465 to the right, improve=1.7561930, (0 missing)
## age < 59.5 to the right, improve=1.2446620, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.3864080, (0 missing)
## stroke < 0.5 to the left, improve=0.3262151, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1237742, (0 missing)
##
## Node number 15447: 40 observations
## predicted class=B1 expected loss=0.65 P(node) =0.0001455588
## class counts: 14 12 12 2 0
## probabilities: 0.350 0.300 0.300 0.050 0.000
##
## Node number 15450: 703 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.598862 P(node) =0.002558196
## class counts: 282 244 119 53 5
## probabilities: 0.401 0.347 0.169 0.075 0.007
## left son=30900 (298 obs) right son=30901 (405 obs)
## Primary splits:
## reimbursement2008 < 6635 to the right, improve=1.9072210, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4867600, (0 missing)
## age < 74.5 to the left, improve=1.1374550, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4608058, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4586126, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.596, adj=0.047, (0 split)
## age < 35.5 to the left, agree=0.578, adj=0.003, (0 split)
##
## Node number 15451: 34 observations
## predicted class=B2 expected loss=0.5 P(node) =0.000123725
## class counts: 8 17 4 5 0
## probabilities: 0.235 0.500 0.118 0.147 0.000
##
## Node number 15452: 297 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5757576 P(node) =0.001080774
## class counts: 126 113 45 12 1
## probabilities: 0.424 0.380 0.152 0.040 0.003
## left son=30904 (274 obs) right son=30905 (23 obs)
## Primary splits:
## reimbursement2008 < 5065 to the left, improve=2.3768610, (0 missing)
## alzheimers < 0.5 to the right, improve=2.2936150, (0 missing)
## age < 37.5 to the left, improve=1.9456180, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.7719678, (0 missing)
## stroke < 0.5 to the left, improve=0.6098027, (0 missing)
##
## Node number 15453: 700 observations
## predicted class=B2 expected loss=0.5814286 P(node) =0.002547279
## class counts: 231 293 126 45 5
## probabilities: 0.330 0.419 0.180 0.064 0.007
##
## Node number 15458: 889 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.003235045
## class counts: 327 381 133 45 3
## probabilities: 0.368 0.429 0.150 0.051 0.003
##
## Node number 15459: 1053 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5318139 P(node) =0.003831836
## class counts: 315 493 175 64 6
## probabilities: 0.299 0.468 0.166 0.061 0.006
## left son=30918 (721 obs) right son=30919 (332 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.319421, (0 missing)
## age < 65.5 to the right, improve=2.157808, (0 missing)
## reimbursement2008 < 4195 to the left, improve=2.005955, (0 missing)
## stroke < 0.5 to the left, improve=1.694776, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.685891, (0 missing)
##
## Node number 15480: 83 observations
## predicted class=B2 expected loss=0.5662651 P(node) =0.0003020345
## class counts: 29 36 8 10 0
## probabilities: 0.349 0.434 0.096 0.120 0.000
##
## Node number 15481: 556 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6330935 P(node) =0.002023268
## class counts: 204 185 116 49 2
## probabilities: 0.367 0.333 0.209 0.088 0.004
## left son=30962 (368 obs) right son=30963 (188 obs)
## Primary splits:
## age < 67.5 to the right, improve=1.7538220, (0 missing)
## reimbursement2008 < 17290 to the right, improve=1.5233210, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8892958, (0 missing)
## stroke < 0.5 to the left, improve=0.8663588, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8033839, (0 missing)
##
## Node number 15484: 40 observations
## predicted class=B1 expected loss=0.425 P(node) =0.0001455588
## class counts: 23 8 7 2 0
## probabilities: 0.575 0.200 0.175 0.050 0.000
##
## Node number 15485: 82 observations
## predicted class=B2 expected loss=0.6219512 P(node) =0.0002983956
## class counts: 26 31 15 10 0
## probabilities: 0.317 0.378 0.183 0.122 0.000
##
## Node number 15884: 93 observations, complexity param=5.811572e-05
## predicted class=B2 expected loss=0.5913978 P(node) =0.0003384243
## class counts: 32 38 15 8 0
## probabilities: 0.344 0.409 0.161 0.086 0.000
## left son=31768 (44 obs) right son=31769 (49 obs)
## Primary splits:
## reimbursement2008 < 6110 to the right, improve=2.8038180, (0 missing)
## age < 68.5 to the right, improve=0.9063337, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4118188, (0 missing)
## alzheimers < 0.5 to the left, improve=0.3578690, (0 missing)
## stroke < 0.5 to the right, improve=0.3151562, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.785, adj=0.545, (0 split)
## age < 73.5 to the left, agree=0.602, adj=0.159, (0 split)
## alzheimers < 0.5 to the right, agree=0.538, adj=0.023, (0 split)
##
## Node number 15885: 13 observations
## predicted class=B1 expected loss=0.4615385 P(node) =4.730662e-05
## class counts: 7 1 1 4 0
## probabilities: 0.538 0.077 0.077 0.308 0.000
##
## Node number 15920: 331 observations
## predicted class=B1 expected loss=0.6344411 P(node) =0.001204499
## class counts: 121 94 53 53 10
## probabilities: 0.366 0.284 0.160 0.160 0.030
##
## Node number 15921: 18 observations
## predicted class=B2 expected loss=0.5 P(node) =6.550147e-05
## class counts: 2 9 2 4 1
## probabilities: 0.111 0.500 0.111 0.222 0.056
##
## Node number 15924: 310 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6741935 P(node) =0.001128081
## class counts: 101 99 64 37 9
## probabilities: 0.326 0.319 0.206 0.119 0.029
## left son=31848 (50 obs) right son=31849 (260 obs)
## Primary splits:
## reimbursement2008 < 9955 to the left, improve=3.5194040, (0 missing)
## alzheimers < 0.5 to the left, improve=1.4052180, (0 missing)
## age < 60.5 to the right, improve=1.3545900, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9417634, (0 missing)
## stroke < 0.5 to the right, improve=0.4401818, (0 missing)
##
## Node number 15925: 237 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6582278 P(node) =0.000862436
## class counts: 53 81 72 28 3
## probabilities: 0.224 0.342 0.304 0.118 0.013
## left son=31850 (56 obs) right son=31851 (181 obs)
## Primary splits:
## age < 67.5 to the left, improve=3.14488400, (0 missing)
## reimbursement2008 < 7130 to the left, improve=2.11196700, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.86604090, (0 missing)
## stroke < 0.5 to the left, improve=0.39390990, (0 missing)
## alzheimers < 0.5 to the right, improve=0.05008339, (0 missing)
##
## Node number 16280: 398 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7236181 P(node) =0.00144831
## class counts: 110 101 63 99 25
## probabilities: 0.276 0.254 0.158 0.249 0.063
## left son=32560 (179 obs) right son=32561 (219 obs)
## Primary splits:
## alzheimers < 0.5 to the right, improve=2.797541, (0 missing)
## ihd < 0.5 to the right, improve=2.182276, (0 missing)
## reimbursement2008 < 15500 to the right, improve=1.710577, (0 missing)
## stroke < 0.5 to the right, improve=1.223226, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.211249, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the right, agree=0.563, adj=0.028, (0 split)
## reimbursement2008 < 15625 to the left, agree=0.563, adj=0.028, (0 split)
## age < 62.5 to the left, agree=0.555, adj=0.011, (0 split)
## osteoporosis < 0.5 to the right, agree=0.553, adj=0.006, (0 split)
##
## Node number 16281: 59 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6440678 P(node) =0.0002146993
## class counts: 21 6 10 21 1
## probabilities: 0.356 0.102 0.169 0.356 0.017
## left son=32562 (18 obs) right son=32563 (41 obs)
## Primary splits:
## reimbursement2008 < 19680 to the right, improve=1.9754260, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9501021, (0 missing)
## bucket2008 < 3.5 to the right, improve=0.8931654, (0 missing)
## age < 90.5 to the right, improve=0.7250257, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5260164, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.847, adj=0.5, (0 split)
##
## Node number 27764: 343 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5568513 P(node) =0.001248167
## class counts: 152 136 37 16 2
## probabilities: 0.443 0.397 0.108 0.047 0.006
## left son=55528 (117 obs) right son=55529 (226 obs)
## Primary splits:
## reimbursement2008 < 2835 to the right, improve=1.9282960, (0 missing)
## stroke < 0.5 to the left, improve=1.1581140, (0 missing)
## age < 82.5 to the right, improve=1.0933820, (0 missing)
## kidney < 0.5 to the left, improve=1.0145490, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9380155, (0 missing)
##
## Node number 27765: 129 observations
## predicted class=B1 expected loss=0.4496124 P(node) =0.0004694272
## class counts: 71 27 20 9 2
## probabilities: 0.550 0.209 0.155 0.070 0.016
##
## Node number 27766: 218 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5733945 P(node) =0.0007932956
## class counts: 93 89 28 8 0
## probabilities: 0.427 0.408 0.128 0.037 0.000
## left son=55532 (194 obs) right son=55533 (24 obs)
## Primary splits:
## reimbursement2008 < 2945 to the left, improve=1.9617420, (0 missing)
## depression < 0.5 to the left, improve=0.6526821, (0 missing)
## kidney < 0.5 to the left, improve=0.4610298, (0 missing)
## age < 57.5 to the left, improve=0.4574831, (0 missing)
## alzheimers < 0.5 to the left, improve=0.3559027, (0 missing)
##
## Node number 27767: 79 observations
## predicted class=B2 expected loss=0.4810127 P(node) =0.0002874787
## class counts: 26 41 7 5 0
## probabilities: 0.329 0.519 0.089 0.063 0.000
##
## Node number 28002: 192 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.59375 P(node) =0.0006986823
## class counts: 72 78 29 11 2
## probabilities: 0.375 0.406 0.151 0.057 0.010
## left son=56004 (124 obs) right son=56005 (68 obs)
## Primary splits:
## age < 78.5 to the left, improve=3.1968100, (0 missing)
## reimbursement2008 < 2885 to the left, improve=2.1236740, (0 missing)
## alzheimers < 0.5 to the right, improve=1.0053880, (0 missing)
## bucket2008 < 1.5 to the left, improve=0.7479369, (0 missing)
## stroke < 0.5 to the left, improve=0.5316513, (0 missing)
##
## Node number 28003: 74 observations, complexity param=7.19528e-05
## predicted class=B3 expected loss=0.6486486 P(node) =0.0002692838
## class counts: 24 22 26 2 0
## probabilities: 0.324 0.297 0.351 0.027 0.000
## left son=56006 (8 obs) right son=56007 (66 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=6.4864860, (0 missing)
## age < 65 to the left, improve=1.7666590, (0 missing)
## alzheimers < 0.5 to the left, improve=1.7622440, (0 missing)
## reimbursement2008 < 2355 to the right, improve=1.0927360, (0 missing)
## stroke < 0.5 to the left, improve=0.8745462, (0 missing)
##
## Node number 28066: 169 observations
## predicted class=B1 expected loss=0.5798817 P(node) =0.000614986
## class counts: 71 58 32 6 2
## probabilities: 0.420 0.343 0.189 0.036 0.012
##
## Node number 28067: 45 observations
## predicted class=B2 expected loss=0.5111111 P(node) =0.0001637537
## class counts: 12 22 9 2 0
## probabilities: 0.267 0.489 0.200 0.044 0.000
##
## Node number 28354: 169 observations
## predicted class=B1 expected loss=0.5621302 P(node) =0.000614986
## class counts: 74 60 26 8 1
## probabilities: 0.438 0.355 0.154 0.047 0.006
##
## Node number 28355: 62 observations
## predicted class=B2 expected loss=0.4516129 P(node) =0.0002256162
## class counts: 20 34 6 2 0
## probabilities: 0.323 0.548 0.097 0.032 0.000
##
## Node number 28358: 689 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5718433 P(node) =0.002507251
## class counts: 295 261 92 38 3
## probabilities: 0.428 0.379 0.134 0.055 0.004
## left son=56716 (367 obs) right son=56717 (322 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=1.5366830, (0 missing)
## reimbursement2008 < 2185 to the right, improve=0.9498001, (0 missing)
## age < 67.5 to the left, improve=0.9450906, (0 missing)
## copd < 0.5 to the left, improve=0.5052370, (0 missing)
## depression < 0.5 to the left, improve=0.4336301, (0 missing)
## Surrogate splits:
## copd < 0.5 to the left, agree=0.605, adj=0.155, (0 split)
## age < 85.5 to the left, agree=0.543, adj=0.022, (0 split)
## reimbursement2008 < 2515 to the left, agree=0.541, adj=0.019, (0 split)
## alzheimers < 0.5 to the left, agree=0.538, adj=0.012, (0 split)
##
## Node number 28359: 34 observations
## predicted class=B2 expected loss=0.5294118 P(node) =0.000123725
## class counts: 8 16 6 4 0
## probabilities: 0.235 0.471 0.176 0.118 0.000
##
## Node number 30808: 253 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5454545 P(node) =0.0009206595
## class counts: 115 89 30 17 2
## probabilities: 0.455 0.352 0.119 0.067 0.008
## left son=61616 (245 obs) right son=61617 (8 obs)
## Primary splits:
## age < 96 to the left, improve=1.6668230, (0 missing)
## reimbursement2008 < 8170 to the left, improve=1.5801570, (0 missing)
## depression < 0.5 to the left, improve=0.8012407, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6406559, (0 missing)
## stroke < 0.5 to the left, improve=0.4810539, (0 missing)
##
## Node number 30809: 56 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0002037823
## class counts: 17 28 8 3 0
## probabilities: 0.304 0.500 0.143 0.054 0.000
##
## Node number 30892: 1478 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5886333 P(node) =0.005378398
## class counts: 600 608 199 67 4
## probabilities: 0.406 0.411 0.135 0.045 0.003
## left son=61784 (759 obs) right son=61785 (719 obs)
## Primary splits:
## reimbursement2008 < 4655 to the left, improve=1.4912330, (0 missing)
## age < 59.5 to the right, improve=1.4379920, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4252592, (0 missing)
## stroke < 0.5 to the left, improve=0.4189515, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.1287486, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.566, adj=0.108, (0 split)
## alzheimers < 0.5 to the left, agree=0.535, adj=0.043, (0 split)
## stroke < 0.5 to the left, agree=0.533, adj=0.040, (0 split)
## copd < 0.5 to the left, agree=0.530, adj=0.033, (0 split)
## age < 82.5 to the left, agree=0.526, adj=0.025, (0 split)
##
## Node number 30893: 49 observations
## predicted class=B2 expected loss=0.4897959 P(node) =0.0001783096
## class counts: 13 25 4 5 2
## probabilities: 0.265 0.510 0.082 0.102 0.041
##
## Node number 30900: 298 observations
## predicted class=B1 expected loss=0.5503356 P(node) =0.001084413
## class counts: 134 94 46 20 4
## probabilities: 0.450 0.315 0.154 0.067 0.013
##
## Node number 30901: 405 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.6296296 P(node) =0.001473783
## class counts: 148 150 73 33 1
## probabilities: 0.365 0.370 0.180 0.081 0.002
## left son=61802 (137 obs) right son=61803 (268 obs)
## Primary splits:
## reimbursement2008 < 5685 to the left, improve=1.4352860, (0 missing)
## age < 43 to the right, improve=1.2563810, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8108852, (0 missing)
## alzheimers < 0.5 to the right, improve=0.3644866, (0 missing)
## copd < 0.5 to the left, improve=0.3421456, (0 missing)
##
## Node number 30904: 274 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5583942 P(node) =0.0009970779
## class counts: 121 99 42 11 1
## probabilities: 0.442 0.361 0.153 0.040 0.004
## left son=61808 (174 obs) right son=61809 (100 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.2349370, (0 missing)
## age < 37.5 to the left, improve=1.7714310, (0 missing)
## reimbursement2008 < 4990 to the right, improve=1.7636660, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8701440, (0 missing)
## stroke < 0.5 to the left, improve=0.4025273, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.668, adj=0.09, (0 split)
## reimbursement2008 < 3085 to the right, agree=0.642, adj=0.02, (0 split)
##
## Node number 30905: 23 observations
## predicted class=B2 expected loss=0.3913043 P(node) =8.369632e-05
## class counts: 5 14 3 1 0
## probabilities: 0.217 0.609 0.130 0.043 0.000
##
## Node number 30918: 721 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5492372 P(node) =0.002623698
## class counts: 234 325 114 44 4
## probabilities: 0.325 0.451 0.158 0.061 0.006
## left son=61836 (109 obs) right son=61837 (612 obs)
## Primary splits:
## age < 86.5 to the right, improve=5.2024390, (0 missing)
## reimbursement2008 < 8105 to the left, improve=1.9497410, (0 missing)
## stroke < 0.5 to the left, improve=1.3441110, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.8592657, (0 missing)
## alzheimers < 0.5 to the left, improve=0.1415473, (0 missing)
##
## Node number 30919: 332 observations
## predicted class=B2 expected loss=0.4939759 P(node) =0.001208138
## class counts: 81 168 61 20 2
## probabilities: 0.244 0.506 0.184 0.060 0.006
##
## Node number 30962: 368 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.625 P(node) =0.001339141
## class counts: 138 132 66 32 0
## probabilities: 0.375 0.359 0.179 0.087 0.000
## left son=61924 (261 obs) right son=61925 (107 obs)
## Primary splits:
## reimbursement2008 < 10440 to the right, improve=2.0386870, (0 missing)
## age < 68.5 to the right, improve=2.0238320, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.0604210, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8507150, (0 missing)
## heart.failure < 0.5 to the right, improve=0.3195541, (0 missing)
##
## Node number 30963: 188 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6489362 P(node) =0.0006841264
## class counts: 66 53 50 17 2
## probabilities: 0.351 0.282 0.266 0.090 0.011
## left son=61926 (135 obs) right son=61927 (53 obs)
## Primary splits:
## age < 55.5 to the right, improve=1.3142350, (0 missing)
## reimbursement2008 < 8995 to the left, improve=1.1323620, (0 missing)
## stroke < 0.5 to the left, improve=0.7672950, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7658279, (0 missing)
## bucket2008 < 3.5 to the right, improve=0.3998270, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8645 to the right, agree=0.723, adj=0.019, (0 split)
##
## Node number 31768: 44 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5227273 P(node) =0.0001601147
## class counts: 21 13 5 5 0
## probabilities: 0.477 0.295 0.114 0.114 0.000
## left son=63536 (26 obs) right son=63537 (18 obs)
## Primary splits:
## reimbursement2008 < 9180 to the left, improve=3.34188000, (0 missing)
## age < 73.5 to the right, improve=1.53473700, (0 missing)
## bucket2008 < 2.5 to the left, improve=1.08333300, (0 missing)
## alzheimers < 0.5 to the left, improve=0.99564270, (0 missing)
## copd < 0.5 to the right, improve=0.09090909, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.864, adj=0.667, (0 split)
## age < 82.5 to the left, agree=0.636, adj=0.111, (0 split)
## stroke < 0.5 to the left, agree=0.614, adj=0.056, (0 split)
##
## Node number 31769: 49 observations
## predicted class=B2 expected loss=0.4897959 P(node) =0.0001783096
## class counts: 11 25 10 3 0
## probabilities: 0.224 0.510 0.204 0.061 0.000
##
## Node number 31848: 50 observations
## predicted class=B2 expected loss=0.56 P(node) =0.0001819485
## class counts: 21 22 1 6 0
## probabilities: 0.420 0.440 0.020 0.120 0.000
##
## Node number 31849: 260 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6923077 P(node) =0.0009461323
## class counts: 80 77 63 31 9
## probabilities: 0.308 0.296 0.242 0.119 0.035
## left son=63698 (20 obs) right son=63699 (240 obs)
## Primary splits:
## reimbursement2008 < 14765 to the right, improve=1.866026, (0 missing)
## alzheimers < 0.5 to the left, improve=1.724179, (0 missing)
## age < 59 to the right, improve=1.389622, (0 missing)
## heart.failure < 0.5 to the left, improve=1.186623, (0 missing)
## stroke < 0.5 to the right, improve=0.396978, (0 missing)
##
## Node number 31850: 56 observations
## predicted class=B2 expected loss=0.5178571 P(node) =0.0002037823
## class counts: 11 27 9 9 0
## probabilities: 0.196 0.482 0.161 0.161 0.000
##
## Node number 31851: 181 observations, complexity param=5.534831e-05
## predicted class=B3 expected loss=0.6519337 P(node) =0.0006586537
## class counts: 42 54 63 19 3
## probabilities: 0.232 0.298 0.348 0.105 0.017
## left son=63702 (136 obs) right son=63703 (45 obs)
## Primary splits:
## reimbursement2008 < 6865 to the right, improve=2.6510090, (0 missing)
## age < 95.5 to the left, improve=1.1712710, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.4758931, (0 missing)
## stroke < 0.5 to the left, improve=0.1841866, (0 missing)
## heart.failure < 0.5 to the left, improve=0.1010412, (0 missing)
##
## Node number 32560: 179 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.6815642 P(node) =0.0006513757
## class counts: 57 51 27 31 13
## probabilities: 0.318 0.285 0.151 0.173 0.073
## left son=65120 (38 obs) right son=65121 (141 obs)
## Primary splits:
## reimbursement2008 < 21440 to the right, improve=2.8400160, (0 missing)
## age < 70.5 to the left, improve=1.0471050, (0 missing)
## stroke < 0.5 to the right, improve=0.8887163, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8119666, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6975642, (0 missing)
##
## Node number 32561: 219 observations
## predicted class=B4 expected loss=0.6894977 P(node) =0.0007969345
## class counts: 53 50 36 68 12
## probabilities: 0.242 0.228 0.164 0.311 0.055
##
## Node number 32562: 18 observations
## predicted class=B1 expected loss=0.4444444 P(node) =6.550147e-05
## class counts: 10 2 0 5 1
## probabilities: 0.556 0.111 0.000 0.278 0.056
##
## Node number 32563: 41 observations
## predicted class=B4 expected loss=0.6097561 P(node) =0.0001491978
## class counts: 11 4 10 16 0
## probabilities: 0.268 0.098 0.244 0.390 0.000
##
## Node number 55528: 117 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4871795 P(node) =0.0004257595
## class counts: 60 38 15 3 1
## probabilities: 0.513 0.325 0.128 0.026 0.009
## left son=111056 (78 obs) right son=111057 (39 obs)
## Primary splits:
## reimbursement2008 < 2945 to the left, improve=2.9829060, (0 missing)
## kidney < 0.5 to the left, improve=1.2210830, (0 missing)
## age < 76.5 to the right, improve=1.1210830, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9103070, (0 missing)
## stroke < 0.5 to the left, improve=0.2543679, (0 missing)
## Surrogate splits:
## age < 76.5 to the right, agree=0.684, adj=0.051, (0 split)
##
## Node number 55529: 226 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5663717 P(node) =0.0008224073
## class counts: 92 98 22 13 1
## probabilities: 0.407 0.434 0.097 0.058 0.004
## left son=111058 (72 obs) right son=111059 (154 obs)
## Primary splits:
## age < 80.5 to the right, improve=1.5914710, (0 missing)
## stroke < 0.5 to the left, improve=1.3555880, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7668454, (0 missing)
## reimbursement2008 < 2795 to the left, improve=0.6874895, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4980787, (0 missing)
##
## Node number 55532: 194 observations
## predicted class=B1 expected loss=0.556701 P(node) =0.0007059603
## class counts: 86 74 27 7 0
## probabilities: 0.443 0.381 0.139 0.036 0.000
##
## Node number 55533: 24 observations
## predicted class=B2 expected loss=0.375 P(node) =8.733529e-05
## class counts: 7 15 1 1 0
## probabilities: 0.292 0.625 0.042 0.042 0.000
##
## Node number 56004: 124 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5403226 P(node) =0.0004512323
## class counts: 57 47 16 4 0
## probabilities: 0.460 0.379 0.129 0.032 0.000
## left son=112008 (46 obs) right son=112009 (78 obs)
## Primary splits:
## age < 72.5 to the right, improve=2.8817380, (0 missing)
## reimbursement2008 < 2885 to the left, improve=1.5254660, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4454760, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7103829, (0 missing)
## stroke < 0.5 to the left, improve=0.4023915, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the right, agree=0.661, adj=0.087, (0 split)
## reimbursement2008 < 2575 to the left, agree=0.645, adj=0.043, (0 split)
##
## Node number 56005: 68 observations
## predicted class=B2 expected loss=0.5441176 P(node) =0.00024745
## class counts: 15 31 13 7 2
## probabilities: 0.221 0.456 0.191 0.103 0.029
##
## Node number 56006: 8 observations
## predicted class=B2 expected loss=0 P(node) =2.911176e-05
## class counts: 0 8 0 0 0
## probabilities: 0.000 1.000 0.000 0.000 0.000
##
## Node number 56007: 66 observations, complexity param=5.534831e-05
## predicted class=B3 expected loss=0.6060606 P(node) =0.0002401721
## class counts: 24 14 26 2 0
## probabilities: 0.364 0.212 0.394 0.030 0.000
## left son=112014 (40 obs) right son=112015 (26 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.3576920, (0 missing)
## age < 65 to the left, improve=1.8352940, (0 missing)
## reimbursement2008 < 2375 to the right, improve=1.1494920, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.1363636, (0 missing)
## Surrogate splits:
## age < 82.5 to the left, agree=0.652, adj=0.115, (0 split)
##
## Node number 56716: 367 observations
## predicted class=B1 expected loss=0.5395095 P(node) =0.001335502
## class counts: 169 131 51 13 3
## probabilities: 0.460 0.357 0.139 0.035 0.008
##
## Node number 56717: 322 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5962733 P(node) =0.001171748
## class counts: 126 130 41 25 0
## probabilities: 0.391 0.404 0.127 0.078 0.000
## left son=113434 (78 obs) right son=113435 (244 obs)
## Primary splits:
## age < 67.5 to the left, improve=2.0124890, (0 missing)
## reimbursement2008 < 2265 to the right, improve=1.1949400, (0 missing)
## alzheimers < 0.5 to the right, improve=0.3273471, (0 missing)
## depression < 0.5 to the right, improve=0.1786959, (0 missing)
## copd < 0.5 to the left, improve=0.1745923, (0 missing)
##
## Node number 61616: 245 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5346939 P(node) =0.0008915478
## class counts: 114 87 27 16 1
## probabilities: 0.465 0.355 0.110 0.065 0.004
## left son=123232 (209 obs) right son=123233 (36 obs)
## Primary splits:
## reimbursement2008 < 8170 to the left, improve=1.7182870, (0 missing)
## age < 90.5 to the right, improve=1.6062760, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.7459219, (0 missing)
## depression < 0.5 to the left, improve=0.6596720, (0 missing)
## stroke < 0.5 to the left, improve=0.6366849, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.971, adj=0.806, (0 split)
##
## Node number 61617: 8 observations
## predicted class=B3 expected loss=0.625 P(node) =2.911176e-05
## class counts: 1 2 3 1 1
## probabilities: 0.125 0.250 0.375 0.125 0.125
##
## Node number 61784: 759 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5678524 P(node) =0.002761979
## class counts: 328 303 94 33 1
## probabilities: 0.432 0.399 0.124 0.043 0.001
## left son=123568 (158 obs) right son=123569 (601 obs)
## Primary splits:
## reimbursement2008 < 4315 to the right, improve=1.62186500, (0 missing)
## age < 82.5 to the right, improve=0.60286370, (0 missing)
## alzheimers < 0.5 to the right, improve=0.24697950, (0 missing)
## copd < 0.5 to the left, improve=0.10233690, (0 missing)
## stroke < 0.5 to the left, improve=0.09394217, (0 missing)
##
## Node number 61785: 719 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5757997 P(node) =0.00261642
## class counts: 272 305 105 34 3
## probabilities: 0.378 0.424 0.146 0.047 0.004
## left son=123570 (346 obs) right son=123571 (373 obs)
## Primary splits:
## reimbursement2008 < 5835 to the left, improve=2.8015510, (0 missing)
## age < 59.5 to the right, improve=2.2849680, (0 missing)
## alzheimers < 0.5 to the right, improve=0.5855315, (0 missing)
## stroke < 0.5 to the right, improve=0.5109046, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2469968, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.590, adj=0.147, (0 split)
## alzheimers < 0.5 to the left, agree=0.537, adj=0.038, (0 split)
## age < 60.5 to the left, agree=0.527, adj=0.017, (0 split)
##
## Node number 61802: 137 observations
## predicted class=B1 expected loss=0.5839416 P(node) =0.000498539
## class counts: 57 43 22 15 0
## probabilities: 0.416 0.314 0.161 0.109 0.000
##
## Node number 61803: 268 observations
## predicted class=B2 expected loss=0.6007463 P(node) =0.0009752441
## class counts: 91 107 51 18 1
## probabilities: 0.340 0.399 0.190 0.067 0.004
##
## Node number 61808: 174 observations
## predicted class=B1 expected loss=0.5229885 P(node) =0.0006331809
## class counts: 83 53 29 8 1
## probabilities: 0.477 0.305 0.167 0.046 0.006
##
## Node number 61809: 100 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.54 P(node) =0.000363897
## class counts: 38 46 13 3 0
## probabilities: 0.380 0.460 0.130 0.030 0.000
## left son=123618 (26 obs) right son=123619 (74 obs)
## Primary splits:
## reimbursement2008 < 4355 to the right, improve=5.3372560, (0 missing)
## age < 62.5 to the right, improve=1.9704690, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.2703500, (0 missing)
## stroke < 0.5 to the left, improve=0.5886275, (0 missing)
## Surrogate splits:
## age < 50.5 to the left, agree=0.79, adj=0.192, (0 split)
##
## Node number 61836: 109 observations
## predicted class=B1 expected loss=0.5412844 P(node) =0.0003966478
## class counts: 50 33 16 9 1
## probabilities: 0.459 0.303 0.147 0.083 0.009
##
## Node number 61837: 612 observations
## predicted class=B2 expected loss=0.5228758 P(node) =0.00222705
## class counts: 184 292 98 35 3
## probabilities: 0.301 0.477 0.160 0.057 0.005
##
## Node number 61924: 261 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5862069 P(node) =0.0009497713
## class counts: 108 87 44 22 0
## probabilities: 0.414 0.333 0.169 0.084 0.000
## left son=123848 (92 obs) right son=123849 (169 obs)
## Primary splits:
## reimbursement2008 < 12585 to the left, improve=2.1315740, (0 missing)
## age < 77.5 to the right, improve=1.2761660, (0 missing)
## stroke < 0.5 to the left, improve=1.0543160, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.0296720, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5202291, (0 missing)
##
## Node number 61925: 107 observations
## predicted class=B2 expected loss=0.5794393 P(node) =0.0003893698
## class counts: 30 45 22 10 0
## probabilities: 0.280 0.421 0.206 0.093 0.000
##
## Node number 61926: 135 observations
## predicted class=B1 expected loss=0.6074074 P(node) =0.000491261
## class counts: 53 34 36 12 0
## probabilities: 0.393 0.252 0.267 0.089 0.000
##
## Node number 61927: 53 observations
## predicted class=B2 expected loss=0.6415094 P(node) =0.0001928654
## class counts: 13 19 14 5 2
## probabilities: 0.245 0.358 0.264 0.094 0.038
##
## Node number 63536: 26 observations
## predicted class=B1 expected loss=0.3461538 P(node) =9.461323e-05
## class counts: 17 4 2 3 0
## probabilities: 0.654 0.154 0.077 0.115 0.000
##
## Node number 63537: 18 observations
## predicted class=B2 expected loss=0.5 P(node) =6.550147e-05
## class counts: 4 9 3 2 0
## probabilities: 0.222 0.500 0.167 0.111 0.000
##
## Node number 63698: 20 observations
## predicted class=B1 expected loss=0.45 P(node) =7.277941e-05
## class counts: 11 5 2 1 1
## probabilities: 0.550 0.250 0.100 0.050 0.050
##
## Node number 63699: 240 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.7 P(node) =0.0008733529
## class counts: 69 72 61 30 8
## probabilities: 0.288 0.300 0.254 0.125 0.033
## left son=127398 (201 obs) right son=127399 (39 obs)
## Primary splits:
## age < 61.5 to the right, improve=1.4580460, (0 missing)
## reimbursement2008 < 10970 to the right, improve=1.4206140, (0 missing)
## alzheimers < 0.5 to the left, improve=1.0755290, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9752886, (0 missing)
## stroke < 0.5 to the right, improve=0.4524283, (0 missing)
##
## Node number 63702: 136 observations
## predicted class=B3 expected loss=0.6029412 P(node) =0.0004949
## class counts: 34 36 54 10 2
## probabilities: 0.250 0.265 0.397 0.074 0.015
##
## Node number 63703: 45 observations
## predicted class=B2 expected loss=0.6 P(node) =0.0001637537
## class counts: 8 18 9 9 1
## probabilities: 0.178 0.400 0.200 0.200 0.022
##
## Node number 65120: 38 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0001382809
## class counts: 9 19 4 5 1
## probabilities: 0.237 0.500 0.105 0.132 0.026
##
## Node number 65121: 141 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6595745 P(node) =0.0005130948
## class counts: 48 32 23 26 12
## probabilities: 0.340 0.227 0.163 0.184 0.085
## left son=130242 (89 obs) right son=130243 (52 obs)
## Primary splits:
## reimbursement2008 < 17585 to the right, improve=2.1889060, (0 missing)
## age < 47.5 to the right, improve=1.2186760, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.1702130, (0 missing)
## stroke < 0.5 to the right, improve=0.9175166, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5919705, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.702, adj=0.192, (0 split)
## age < 47.5 to the right, agree=0.667, adj=0.096, (0 split)
##
## Node number 111056: 78 observations
## predicted class=B1 expected loss=0.4102564 P(node) =0.0002838397
## class counts: 46 19 11 2 0
## probabilities: 0.590 0.244 0.141 0.026 0.000
##
## Node number 111057: 39 observations
## predicted class=B2 expected loss=0.5128205 P(node) =0.0001419198
## class counts: 14 19 4 1 1
## probabilities: 0.359 0.487 0.103 0.026 0.026
##
## Node number 111058: 72 observations
## predicted class=B1 expected loss=0.4861111 P(node) =0.0002620059
## class counts: 37 29 5 1 0
## probabilities: 0.514 0.403 0.069 0.014 0.000
##
## Node number 111059: 154 observations
## predicted class=B2 expected loss=0.5519481 P(node) =0.0005604015
## class counts: 55 69 17 12 1
## probabilities: 0.357 0.448 0.110 0.078 0.006
##
## Node number 112008: 46 observations
## predicted class=B1 expected loss=0.4347826 P(node) =0.0001673926
## class counts: 26 10 8 2 0
## probabilities: 0.565 0.217 0.174 0.043 0.000
##
## Node number 112009: 78 observations
## predicted class=B2 expected loss=0.525641 P(node) =0.0002838397
## class counts: 31 37 8 2 0
## probabilities: 0.397 0.474 0.103 0.026 0.000
##
## Node number 112014: 40 observations
## predicted class=B1 expected loss=0.6 P(node) =0.0001455588
## class counts: 16 12 11 1 0
## probabilities: 0.400 0.300 0.275 0.025 0.000
##
## Node number 112015: 26 observations
## predicted class=B3 expected loss=0.4230769 P(node) =9.461323e-05
## class counts: 8 2 15 1 0
## probabilities: 0.308 0.077 0.577 0.038 0.000
##
## Node number 113434: 78 observations
## predicted class=B1 expected loss=0.5512821 P(node) =0.0002838397
## class counts: 35 23 15 5 0
## probabilities: 0.449 0.295 0.192 0.064 0.000
##
## Node number 113435: 244 observations
## predicted class=B2 expected loss=0.5614754 P(node) =0.0008879088
## class counts: 91 107 26 20 0
## probabilities: 0.373 0.439 0.107 0.082 0.000
##
## Node number 123232: 209 observations
## predicted class=B1 expected loss=0.507177 P(node) =0.0007605448
## class counts: 103 70 22 14 0
## probabilities: 0.493 0.335 0.105 0.067 0.000
##
## Node number 123233: 36 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5277778 P(node) =0.0001310029
## class counts: 11 17 5 2 1
## probabilities: 0.306 0.472 0.139 0.056 0.028
## left son=246466 (15 obs) right son=246467 (21 obs)
## Primary splits:
## age < 74.5 to the left, improve=5.3968250, (0 missing)
## reimbursement2008 < 8705 to the right, improve=1.5053320, (0 missing)
## copd < 0.5 to the right, improve=0.3703704, (0 missing)
## depression < 0.5 to the right, improve=0.3527778, (0 missing)
## heart.failure < 0.5 to the left, improve=0.2972583, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8460 to the left, agree=0.611, adj=0.067, (0 split)
##
## Node number 123568: 158 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0005749573
## class counts: 79 55 15 8 1
## probabilities: 0.500 0.348 0.095 0.051 0.006
##
## Node number 123569: 601 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5856905 P(node) =0.002187021
## class counts: 249 248 79 25 0
## probabilities: 0.414 0.413 0.131 0.042 0.000
## left son=247138 (592 obs) right son=247139 (9 obs)
## Primary splits:
## reimbursement2008 < 4295 to the left, improve=3.0859230, (0 missing)
## age < 62.5 to the right, improve=0.8258296, (0 missing)
## copd < 0.5 to the left, improve=0.2730143, (0 missing)
## stroke < 0.5 to the left, improve=0.1209200, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1102521, (0 missing)
##
## Node number 123570: 346 observations
## predicted class=B2 expected loss=0.5202312 P(node) =0.001259084
## class counts: 122 166 44 13 1
## probabilities: 0.353 0.480 0.127 0.038 0.003
##
## Node number 123571: 373 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.5978552 P(node) =0.001357336
## class counts: 150 139 61 21 2
## probabilities: 0.402 0.373 0.164 0.056 0.005
## left son=247142 (124 obs) right son=247143 (249 obs)
## Primary splits:
## alzheimers < 0.5 to the right, improve=1.9370400, (0 missing)
## reimbursement2008 < 6045 to the right, improve=1.9317030, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.1351660, (0 missing)
## age < 68.5 to the left, improve=0.9923350, (0 missing)
## stroke < 0.5 to the right, improve=0.8206414, (0 missing)
## Surrogate splits:
## age < 64.5 to the left, agree=0.673, adj=0.016, (0 split)
## stroke < 0.5 to the right, agree=0.673, adj=0.016, (0 split)
## reimbursement2008 < 5845 to the left, agree=0.670, adj=0.008, (0 split)
##
## Node number 123618: 26 observations
## predicted class=B1 expected loss=0.3846154 P(node) =9.461323e-05
## class counts: 16 4 4 2 0
## probabilities: 0.615 0.154 0.154 0.077 0.000
##
## Node number 123619: 74 observations
## predicted class=B2 expected loss=0.4324324 P(node) =0.0002692838
## class counts: 22 42 9 1 0
## probabilities: 0.297 0.568 0.122 0.014 0.000
##
## Node number 123848: 92 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0003347853
## class counts: 46 23 17 6 0
## probabilities: 0.500 0.250 0.185 0.065 0.000
##
## Node number 123849: 169 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6213018 P(node) =0.000614986
## class counts: 62 64 27 16 0
## probabilities: 0.367 0.379 0.160 0.095 0.000
## left son=247698 (109 obs) right son=247699 (60 obs)
## Primary splits:
## reimbursement2008 < 14485 to the right, improve=2.3703890, (0 missing)
## age < 77.5 to the right, improve=1.8205180, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.5605270, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9473954, (0 missing)
## stroke < 0.5 to the right, improve=0.8779250, (0 missing)
##
## Node number 127398: 201 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6865672 P(node) =0.0007314331
## class counts: 63 58 49 27 4
## probabilities: 0.313 0.289 0.244 0.134 0.020
## left son=254796 (112 obs) right son=254797 (89 obs)
## Primary splits:
## reimbursement2008 < 12625 to the left, improve=2.6465710, (0 missing)
## age < 72.5 to the right, improve=1.5701210, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5204340, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8281641, (0 missing)
## stroke < 0.5 to the right, improve=0.4454147, (0 missing)
## Surrogate splits:
## age < 67.5 to the right, agree=0.587, adj=0.067, (0 split)
##
## Node number 127399: 39 observations
## predicted class=B2 expected loss=0.6410256 P(node) =0.0001419198
## class counts: 6 14 12 3 4
## probabilities: 0.154 0.359 0.308 0.077 0.103
##
## Node number 130242: 89 observations
## predicted class=B1 expected loss=0.5955056 P(node) =0.0003238684
## class counts: 36 15 17 14 7
## probabilities: 0.404 0.169 0.191 0.157 0.079
##
## Node number 130243: 52 observations
## predicted class=B2 expected loss=0.6730769 P(node) =0.0001892265
## class counts: 12 17 6 12 5
## probabilities: 0.231 0.327 0.115 0.231 0.096
##
## Node number 246466: 15 observations
## predicted class=B1 expected loss=0.4 P(node) =5.458456e-05
## class counts: 9 2 3 0 1
## probabilities: 0.600 0.133 0.200 0.000 0.067
##
## Node number 246467: 21 observations
## predicted class=B2 expected loss=0.2857143 P(node) =7.641838e-05
## class counts: 2 15 2 2 0
## probabilities: 0.095 0.714 0.095 0.095 0.000
##
## Node number 247138: 592 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5810811 P(node) =0.002154271
## class counts: 248 240 79 25 0
## probabilities: 0.419 0.405 0.133 0.042 0.000
## left son=494276 (135 obs) right son=494277 (457 obs)
## Primary splits:
## age < 82.5 to the right, improve=1.0162580, (0 missing)
## reimbursement2008 < 3485 to the left, improve=0.9533819, (0 missing)
## copd < 0.5 to the left, improve=0.2603666, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1489946, (0 missing)
## stroke < 0.5 to the left, improve=0.1384892, (0 missing)
##
## Node number 247139: 9 observations
## predicted class=B2 expected loss=0.1111111 P(node) =3.275073e-05
## class counts: 1 8 0 0 0
## probabilities: 0.111 0.889 0.000 0.000 0.000
##
## Node number 247142: 124 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.516129 P(node) =0.0004512323
## class counts: 60 39 19 5 1
## probabilities: 0.484 0.315 0.153 0.040 0.008
## left son=494284 (114 obs) right son=494285 (10 obs)
## Primary splits:
## reimbursement2008 < 8555 to the left, improve=3.2894170, (0 missing)
## age < 62.5 to the right, improve=1.3134040, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8306452, (0 missing)
## stroke < 0.5 to the right, improve=0.6624062, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6169185, (0 missing)
##
## Node number 247143: 249 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5983936 P(node) =0.0009061036
## class counts: 90 100 42 16 1
## probabilities: 0.361 0.402 0.169 0.064 0.004
## left son=494286 (217 obs) right son=494287 (32 obs)
## Primary splits:
## reimbursement2008 < 6045 to the right, improve=2.8382200, (0 missing)
## age < 68.5 to the left, improve=1.5757780, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8580882, (0 missing)
## copd < 0.5 to the left, improve=0.4427711, (0 missing)
## stroke < 0.5 to the right, improve=0.2244234, (0 missing)
##
## Node number 247698: 109 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5779817 P(node) =0.0003966478
## class counts: 46 34 19 10 0
## probabilities: 0.422 0.312 0.174 0.092 0.000
## left son=495396 (72 obs) right son=495397 (37 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.1824740, (0 missing)
## reimbursement2008 < 17235 to the right, improve=1.3957040, (0 missing)
## age < 70.5 to the left, improve=1.2827700, (0 missing)
## stroke < 0.5 to the left, improve=1.2406940, (0 missing)
## bucket2008 < 3.5 to the left, improve=0.2455781, (0 missing)
## Surrogate splits:
## reimbursement2008 < 14660 to the right, agree=0.679, adj=0.054, (0 split)
##
## Node number 247699: 60 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0002183382
## class counts: 16 30 8 6 0
## probabilities: 0.267 0.500 0.133 0.100 0.000
##
## Node number 254796: 112 observations
## predicted class=B1 expected loss=0.6160714 P(node) =0.0004075647
## class counts: 43 27 31 10 1
## probabilities: 0.384 0.241 0.277 0.089 0.009
##
## Node number 254797: 89 observations
## predicted class=B2 expected loss=0.6516854 P(node) =0.0003238684
## class counts: 20 31 18 17 3
## probabilities: 0.225 0.348 0.202 0.191 0.034
##
## Node number 494276: 135 observations
## predicted class=B1 expected loss=0.5259259 P(node) =0.000491261
## class counts: 64 49 20 2 0
## probabilities: 0.474 0.363 0.148 0.015 0.000
##
## Node number 494277: 457 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5820569 P(node) =0.00166301
## class counts: 184 191 59 23 0
## probabilities: 0.403 0.418 0.129 0.050 0.000
## left son=988554 (290 obs) right son=988555 (167 obs)
## Primary splits:
## age < 74.5 to the left, improve=0.9874503, (0 missing)
## reimbursement2008 < 3495 to the left, improve=0.9861916, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3272674, (0 missing)
## stroke < 0.5 to the left, improve=0.2337493, (0 missing)
## copd < 0.5 to the left, improve=0.1994562, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.637, adj=0.006, (0 split)
##
## Node number 494284: 114 observations
## predicted class=B1 expected loss=0.4824561 P(node) =0.0004148426
## class counts: 59 32 18 4 1
## probabilities: 0.518 0.281 0.158 0.035 0.009
##
## Node number 494285: 10 observations
## predicted class=B2 expected loss=0.3 P(node) =3.63897e-05
## class counts: 1 7 1 1 0
## probabilities: 0.100 0.700 0.100 0.100 0.000
##
## Node number 494286: 217 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.0007896566
## class counts: 78 93 30 15 1
## probabilities: 0.359 0.429 0.138 0.069 0.005
##
## Node number 494287: 32 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.625 P(node) =0.0001164471
## class counts: 12 7 12 1 0
## probabilities: 0.375 0.219 0.375 0.031 0.000
## left son=988574 (11 obs) right son=988575 (21 obs)
## Primary splits:
## age < 72.5 to the left, improve=1.8097940, (0 missing)
## reimbursement2008 < 5975 to the left, improve=0.7232143, (0 missing)
## copd < 0.5 to the left, improve=0.6875000, (0 missing)
##
## Node number 495396: 72 observations
## predicted class=B1 expected loss=0.5138889 P(node) =0.0002620059
## class counts: 35 17 12 8 0
## probabilities: 0.486 0.236 0.167 0.111 0.000
##
## Node number 495397: 37 observations
## predicted class=B2 expected loss=0.5405405 P(node) =0.0001346419
## class counts: 11 17 7 2 0
## probabilities: 0.297 0.459 0.189 0.054 0.000
##
## Node number 988554: 290 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5724138 P(node) =0.001055301
## class counts: 124 114 37 15 0
## probabilities: 0.428 0.393 0.128 0.052 0.000
## left son=1977108 (234 obs) right son=1977109 (56 obs)
## Primary splits:
## age < 62.5 to the right, improve=1.0825800, (0 missing)
## reimbursement2008 < 3945 to the right, improve=0.7040408, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.6026089, (0 missing)
## stroke < 0.5 to the left, improve=0.2655768, (0 missing)
## copd < 0.5 to the left, improve=0.1804923, (0 missing)
##
## Node number 988555: 167 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5389222 P(node) =0.0006077081
## class counts: 60 77 22 8 0
## probabilities: 0.359 0.461 0.132 0.048 0.000
## left son=1977110 (39 obs) right son=1977111 (128 obs)
## Primary splits:
## reimbursement2008 < 4105 to the right, improve=1.3886510, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7439669, (0 missing)
## age < 81.5 to the left, improve=0.4824922, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2060442, (0 missing)
## copd < 0.5 to the left, improve=0.1297289, (0 missing)
##
## Node number 988574: 11 observations
## predicted class=B1 expected loss=0.3636364 P(node) =4.002868e-05
## class counts: 7 2 2 0 0
## probabilities: 0.636 0.182 0.182 0.000 0.000
##
## Node number 988575: 21 observations
## predicted class=B3 expected loss=0.5238095 P(node) =7.641838e-05
## class counts: 5 5 10 1 0
## probabilities: 0.238 0.238 0.476 0.048 0.000
##
## Node number 1977108: 234 observations
## predicted class=B1 expected loss=0.5470085 P(node) =0.0008515191
## class counts: 106 89 28 11 0
## probabilities: 0.453 0.380 0.120 0.047 0.000
##
## Node number 1977109: 56 observations
## predicted class=B2 expected loss=0.5535714 P(node) =0.0002037823
## class counts: 18 25 9 4 0
## probabilities: 0.321 0.446 0.161 0.071 0.000
##
## Node number 1977110: 39 observations
## predicted class=B1 expected loss=0.5128205 P(node) =0.0001419198
## class counts: 19 14 5 1 0
## probabilities: 0.487 0.359 0.128 0.026 0.000
##
## Node number 1977111: 128 observations
## predicted class=B2 expected loss=0.5078125 P(node) =0.0004657882
## class counts: 41 63 17 7 0
## probabilities: 0.320 0.492 0.133 0.055 0.000
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) reimbursement2008< 1565 165987 20938 B1 (0.87 0.074 0.037 0.014 0.0014) *
## 3) reimbursement2008>=1565 108816 68841 B2 (0.36 0.37 0.17 0.088 0.012)
## 6) reimbursement2008< 3065 39298 18853 B1 (0.52 0.31 0.12 0.045 0.0046)
## 12) reimbursement2008< 2175 20077 8527 B1 (0.58 0.27 0.11 0.042 0.0038)
## 24) diabetes< 0.5 8826 3280 B1 (0.63 0.24 0.091 0.035 0.0029) *
## 25) diabetes>=0.5 11251 5247 B1 (0.53 0.29 0.12 0.046 0.0045)
## 50) kidney< 0.5 9007 4045 B1 (0.55 0.28 0.12 0.042 0.0044)
## 100) reimbursement2008< 1875 4935 2071 B1 (0.58 0.26 0.11 0.042 0.0045) *
## 101) reimbursement2008>=1875 4072 1974 B1 (0.52 0.31 0.13 0.042 0.0044)
## 202) cancer< 0.5 3786 1807 B1 (0.52 0.3 0.13 0.041 0.0048) *
## 203) cancer>=0.5 286 167 B1 (0.42 0.4 0.13 0.056 0)
## 406) age< 73.5 128 64 B1 (0.5 0.33 0.11 0.062 0)
## 812) depression< 0.5 95 41 B1 (0.57 0.26 0.12 0.053 0) *
## 813) depression>=0.5 33 16 B2 (0.3 0.52 0.091 0.091 0) *
## 407) age>=73.5 158 85 B2 (0.35 0.46 0.14 0.051 0) *
## 51) kidney>=0.5 2244 1202 B1 (0.46 0.33 0.14 0.064 0.0049)
## 102) heart.failure< 0.5 992 477 B1 (0.52 0.29 0.13 0.057 0.002) *
## 103) heart.failure>=0.5 1252 725 B1 (0.42 0.36 0.15 0.069 0.0072)
## 206) arthritis< 0.5 904 486 B1 (0.46 0.34 0.13 0.063 0.0066)
## 412) reimbursement2008< 1735 270 125 B1 (0.54 0.27 0.13 0.056 0.0037) *
## 413) reimbursement2008>=1735 634 361 B1 (0.43 0.36 0.13 0.066 0.0079)
## 826) age< 91.5 596 330 B1 (0.45 0.36 0.13 0.057 0.0084)
## 1652) reimbursement2008>=1765 555 302 B1 (0.46 0.35 0.14 0.052 0.0072)
## 3304) reimbursement2008< 1955 265 130 B1 (0.51 0.31 0.13 0.049 0.0038)
## 6608) stroke< 0.5 251 118 B1 (0.53 0.29 0.13 0.044 0.004)
## 13216) cancer< 0.5 235 106 B1 (0.55 0.27 0.13 0.047 0.0043) *
## 13217) cancer>=0.5 16 7 B2 (0.25 0.56 0.19 0 0) *
## 6609) stroke>=0.5 14 5 B2 (0.14 0.64 0.071 0.14 0) *
## 3305) reimbursement2008>=1955 290 172 B1 (0.41 0.38 0.14 0.055 0.01)
## 6610) age< 81.5 213 120 B1 (0.44 0.35 0.14 0.07 0.0047)
## 13220) age>=44.5 201 110 B1 (0.45 0.33 0.14 0.07 0) *
## 13221) age< 44.5 12 5 B2 (0.17 0.58 0.083 0.083 0.083) *
## 6611) age>=81.5 77 40 B2 (0.32 0.48 0.16 0.013 0.026) *
## 1653) reimbursement2008< 1765 41 20 B2 (0.32 0.51 0.024 0.12 0.024) *
## 827) age>=91.5 38 21 B2 (0.18 0.45 0.16 0.21 0) *
## 207) arthritis>=0.5 348 205 B2 (0.31 0.41 0.18 0.086 0.0086) *
## 13) reimbursement2008>=2175 19221 10326 B1 (0.46 0.35 0.13 0.049 0.0054)
## 26) diabetes< 0.5 7137 3360 B1 (0.53 0.31 0.11 0.042 0.0046)
## 52) arthritis< 0.5 5554 2471 B1 (0.56 0.3 0.1 0.039 0.0049)
## 104) ihd< 0.5 2348 933 B1 (0.6 0.27 0.092 0.031 0.0051) *
## 105) ihd>=0.5 3206 1538 B1 (0.52 0.32 0.11 0.045 0.0047)
## 210) depression< 0.5 2325 1056 B1 (0.55 0.3 0.11 0.043 0.0052) *
## 211) depression>=0.5 881 482 B1 (0.45 0.36 0.13 0.052 0.0034)
## 422) kidney< 0.5 763 405 B1 (0.47 0.34 0.13 0.052 0.0039) *
## 423) kidney>=0.5 118 63 B2 (0.35 0.47 0.14 0.051 0)
## 846) reimbursement2008>=2865 22 10 B1 (0.55 0.23 0.18 0.045 0) *
## 847) reimbursement2008< 2865 96 46 B2 (0.3 0.52 0.12 0.052 0) *
## 53) arthritis>=0.5 1583 889 B1 (0.44 0.37 0.14 0.052 0.0038)
## 106) stroke< 0.5 1525 844 B1 (0.45 0.36 0.13 0.054 0.0039)
## 212) cancer< 0.5 1438 784 B1 (0.45 0.36 0.13 0.053 0.0042)
## 424) reimbursement2008>=2715 495 280 B1 (0.43 0.41 0.1 0.053 0.004)
## 848) reimbursement2008>=2795 385 210 B1 (0.45 0.38 0.1 0.06 0.0052)
## 1696) age< 80.5 263 131 B1 (0.5 0.36 0.099 0.042 0) *
## 1697) age>=80.5 122 70 B2 (0.35 0.43 0.11 0.098 0.016) *
## 849) reimbursement2008< 2795 110 54 B2 (0.36 0.51 0.1 0.027 0) *
## 425) reimbursement2008< 2715 943 504 B1 (0.47 0.33 0.15 0.053 0.0042) *
## 213) cancer>=0.5 87 48 B2 (0.31 0.45 0.17 0.069 0) *
## 107) stroke>=0.5 58 26 B2 (0.22 0.55 0.21 0.017 0) *
## 27) diabetes>=0.5 12084 6966 B1 (0.42 0.37 0.15 0.054 0.0059)
## 54) arthritis< 0.5 8413 4653 B1 (0.45 0.35 0.15 0.052 0.0056)
## 108) heart.failure< 0.5 4375 2220 B1 (0.49 0.34 0.13 0.039 0.0039)
## 216) cancer< 0.5 3992 1978 B1 (0.5 0.33 0.12 0.038 0.0033)
## 432) ihd< 0.5 1265 562 B1 (0.56 0.29 0.12 0.035 0.0032) *
## 433) ihd>=0.5 2727 1416 B1 (0.48 0.35 0.13 0.04 0.0033)
## 866) reimbursement2008< 2615 1499 736 B1 (0.51 0.33 0.13 0.035 0.002) *
## 867) reimbursement2008>=2615 1228 680 B1 (0.45 0.37 0.13 0.046 0.0049)
## 1734) reimbursement2008>=2995 171 81 B1 (0.53 0.27 0.14 0.058 0.0058) *
## 1735) reimbursement2008< 2995 1057 599 B1 (0.43 0.39 0.13 0.044 0.0047)
## 3470) age< 83.5 840 458 B1 (0.45 0.37 0.12 0.049 0.006)
## 6940) age< 54.5 71 31 B1 (0.56 0.23 0.15 0.042 0.014) *
## 6941) age>=54.5 769 427 B1 (0.44 0.38 0.12 0.049 0.0052)
## 13882) age>=70.5 472 249 B1 (0.47 0.35 0.12 0.053 0.0085)
## 27764) age>=73.5 343 191 B1 (0.44 0.4 0.11 0.047 0.0058)
## 55528) reimbursement2008>=2835 117 57 B1 (0.51 0.32 0.13 0.026 0.0085)
## 111056) reimbursement2008< 2945 78 32 B1 (0.59 0.24 0.14 0.026 0) *
## 111057) reimbursement2008>=2945 39 20 B2 (0.36 0.49 0.1 0.026 0.026) *
## 55529) reimbursement2008< 2835 226 128 B2 (0.41 0.43 0.097 0.058 0.0044)
## 111058) age>=80.5 72 35 B1 (0.51 0.4 0.069 0.014 0) *
## 111059) age< 80.5 154 85 B2 (0.36 0.45 0.11 0.078 0.0065) *
## 27765) age< 73.5 129 58 B1 (0.55 0.21 0.16 0.07 0.016) *
## 13883) age< 70.5 297 167 B2 (0.4 0.44 0.12 0.044 0)
## 27766) osteoporosis< 0.5 218 125 B1 (0.43 0.41 0.13 0.037 0)
## 55532) reimbursement2008< 2945 194 108 B1 (0.44 0.38 0.14 0.036 0) *
## 55533) reimbursement2008>=2945 24 9 B2 (0.29 0.62 0.042 0.042 0) *
## 27767) osteoporosis>=0.5 79 38 B2 (0.33 0.52 0.089 0.063 0) *
## 3471) age>=83.5 217 116 B2 (0.35 0.47 0.16 0.028 0) *
## 217) cancer>=0.5 383 220 B2 (0.37 0.43 0.15 0.044 0.01)
## 434) reimbursement2008< 2705 238 136 B1 (0.43 0.36 0.16 0.038 0.013)
## 868) depression< 0.5 167 84 B1 (0.5 0.3 0.15 0.042 0.012) *
## 869) depression>=0.5 71 35 B2 (0.27 0.51 0.18 0.028 0.014) *
## 435) reimbursement2008>=2705 145 68 B2 (0.27 0.53 0.14 0.055 0.0069) *
## 109) heart.failure>=0.5 4038 2433 B1 (0.4 0.36 0.17 0.066 0.0074)
## 218) kidney< 0.5 2819 1620 B1 (0.43 0.35 0.16 0.065 0.0064)
## 436) ihd< 0.5 635 319 B1 (0.5 0.31 0.15 0.041 0.0063) *
## 437) ihd>=0.5 2184 1301 B1 (0.4 0.36 0.16 0.072 0.0064)
## 874) reimbursement2008< 2315 393 202 B1 (0.49 0.34 0.12 0.051 0.0051) *
## 875) reimbursement2008>=2315 1791 1099 B1 (0.39 0.36 0.17 0.076 0.0067)
## 1750) age>=39.5 1752 1066 B1 (0.39 0.36 0.17 0.075 0.0068)
## 3500) depression< 0.5 1099 639 B1 (0.42 0.35 0.15 0.069 0.0064)
## 7000) age< 95.5 1074 620 B1 (0.42 0.35 0.15 0.069 0.0065)
## 14000) copd< 0.5 808 450 B1 (0.44 0.34 0.14 0.075 0.0062) *
## 14001) copd>=0.5 266 166 B2 (0.36 0.38 0.21 0.049 0.0075)
## 28002) reimbursement2008>=2540 192 114 B2 (0.38 0.41 0.15 0.057 0.01)
## 56004) age< 78.5 124 67 B1 (0.46 0.38 0.13 0.032 0)
## 112008) age>=72.5 46 20 B1 (0.57 0.22 0.17 0.043 0) *
## 112009) age< 72.5 78 41 B2 (0.4 0.47 0.1 0.026 0) *
## 56005) age>=78.5 68 37 B2 (0.22 0.46 0.19 0.1 0.029) *
## 28003) reimbursement2008< 2540 74 48 B3 (0.32 0.3 0.35 0.027 0)
## 56006) cancer>=0.5 8 0 B2 (0 1 0 0 0) *
## 56007) cancer< 0.5 66 40 B3 (0.36 0.21 0.39 0.03 0)
## 112014) alzheimers< 0.5 40 24 B1 (0.4 0.3 0.27 0.025 0) *
## 112015) alzheimers>=0.5 26 11 B3 (0.31 0.077 0.58 0.038 0) *
## 7001) age>=95.5 25 10 B2 (0.24 0.6 0.08 0.08 0) *
## 3501) depression>=0.5 653 412 B2 (0.35 0.37 0.19 0.084 0.0077)
## 7002) reimbursement2008< 2655 303 183 B1 (0.4 0.33 0.2 0.069 0.0033) *
## 7003) reimbursement2008>=2655 350 208 B2 (0.3 0.41 0.18 0.097 0.011) *
## 1751) age< 39.5 39 18 B2 (0.15 0.54 0.15 0.15 0) *
## 219) kidney>=0.5 1219 734 B2 (0.33 0.4 0.19 0.07 0.0098)
## 438) reimbursement2008< 2615 613 379 B1 (0.38 0.37 0.18 0.059 0.0098)
## 876) osteoporosis>=0.5 180 102 B1 (0.43 0.38 0.13 0.061 0)
## 1752) reimbursement2008< 2455 112 56 B1 (0.5 0.29 0.12 0.08 0) *
## 1753) reimbursement2008>=2455 68 33 B2 (0.32 0.51 0.13 0.029 0) *
## 877) osteoporosis< 0.5 433 275 B2 (0.36 0.36 0.2 0.058 0.014)
## 1754) stroke< 0.5 403 252 B1 (0.37 0.36 0.2 0.05 0.015)
## 3508) reimbursement2008< 2585 382 238 B2 (0.37 0.38 0.19 0.047 0.01)
## 7016) depression< 0.5 229 136 B1 (0.41 0.36 0.19 0.035 0.0087)
## 14032) cancer>=0.5 15 5 B1 (0.67 0.13 0.2 0 0) *
## 14033) cancer< 0.5 214 131 B1 (0.39 0.37 0.19 0.037 0.0093)
## 28066) reimbursement2008< 2515 169 98 B1 (0.42 0.34 0.19 0.036 0.012) *
## 28067) reimbursement2008>=2515 45 23 B2 (0.27 0.49 0.2 0.044 0) *
## 7017) depression>=0.5 153 91 B2 (0.32 0.41 0.2 0.065 0.013)
## 14034) reimbursement2008>=2545 14 5 B1 (0.64 0.14 0.14 0.071 0) *
## 14035) reimbursement2008< 2545 139 79 B2 (0.29 0.43 0.2 0.065 0.014) *
## 3509) reimbursement2008>=2585 21 12 B1 (0.43 0.14 0.24 0.095 0.095) *
## 1755) stroke>=0.5 30 19 B2 (0.17 0.37 0.3 0.17 0) *
## 439) reimbursement2008>=2615 606 347 B2 (0.28 0.43 0.2 0.081 0.0099) *
## 55) arthritis>=0.5 3671 2129 B2 (0.37 0.42 0.15 0.057 0.0065)
## 110) reimbursement2008< 2665 2068 1224 B1 (0.41 0.4 0.14 0.057 0.0048)
## 220) ihd< 0.5 517 274 B1 (0.47 0.37 0.11 0.048 0.0019)
## 440) reimbursement2008< 2295 143 57 B1 (0.6 0.26 0.077 0.063 0) *
## 441) reimbursement2008>=2295 374 217 B1 (0.42 0.41 0.12 0.043 0.0027)
## 882) reimbursement2008< 2315 25 6 B2 (0.24 0.76 0 0 0) *
## 883) reimbursement2008>=2315 349 198 B1 (0.43 0.39 0.13 0.046 0.0029)
## 1766) cancer< 0.5 336 186 B1 (0.45 0.38 0.13 0.042 0)
## 3532) age< 90.5 322 176 B1 (0.45 0.37 0.13 0.043 0) *
## 3533) age>=90.5 14 5 B2 (0.29 0.64 0.071 0 0) *
## 1767) cancer>=0.5 13 6 B2 (0.077 0.54 0.15 0.15 0.077) *
## 221) ihd>=0.5 1551 925 B2 (0.39 0.4 0.14 0.059 0.0058)
## 442) age< 35 18 5 B1 (0.72 0.22 0 0.056 0) *
## 443) age>=35 1533 911 B2 (0.38 0.41 0.15 0.059 0.0059)
## 886) kidney< 0.5 1101 656 B1 (0.4 0.4 0.14 0.052 0.0045)
## 1772) stroke< 0.5 1057 623 B1 (0.41 0.4 0.14 0.051 0.0038)
## 3544) cancer< 0.5 1008 590 B1 (0.41 0.4 0.13 0.053 0.004)
## 7088) reimbursement2008>=2535 275 154 B2 (0.39 0.44 0.12 0.04 0.0036)
## 14176) age< 63.5 44 17 B2 (0.32 0.61 0.045 0.023 0) *
## 14177) age>=63.5 231 137 B1 (0.41 0.41 0.14 0.043 0.0043)
## 28354) alzheimers< 0.5 169 95 B1 (0.44 0.36 0.15 0.047 0.0059) *
## 28355) alzheimers>=0.5 62 28 B2 (0.32 0.55 0.097 0.032 0) *
## 7089) reimbursement2008< 2535 733 423 B1 (0.42 0.38 0.14 0.057 0.0041)
## 14178) age>=97.5 10 3 B1 (0.7 0.2 0.1 0 0) *
## 14179) age< 97.5 723 420 B1 (0.42 0.38 0.14 0.058 0.0041)
## 28358) age< 90.5 689 394 B1 (0.43 0.38 0.13 0.055 0.0044)
## 56716) heart.failure< 0.5 367 198 B1 (0.46 0.36 0.14 0.035 0.0082) *
## 56717) heart.failure>=0.5 322 192 B2 (0.39 0.4 0.13 0.078 0)
## 113434) age< 67.5 78 43 B1 (0.45 0.29 0.19 0.064 0) *
## 113435) age>=67.5 244 137 B2 (0.37 0.44 0.11 0.082 0) *
## 28359) age>=90.5 34 18 B2 (0.24 0.47 0.18 0.12 0) *
## 3545) cancer>=0.5 49 29 B2 (0.33 0.41 0.24 0.02 0) *
## 1773) stroke>=0.5 44 20 B2 (0.25 0.55 0.11 0.068 0.023) *
## 887) kidney>=0.5 432 254 B2 (0.33 0.41 0.17 0.079 0.0093)
## 1774) reimbursement2008>=2215 403 232 B2 (0.32 0.42 0.17 0.069 0.0099) *
## 1775) reimbursement2008< 2215 29 16 B1 (0.45 0.24 0.1 0.21 0) *
## 111) reimbursement2008>=2665 1603 878 B2 (0.32 0.45 0.16 0.058 0.0087) *
## 7) reimbursement2008>=3065 69518 41677 B2 (0.27 0.4 0.2 0.11 0.017)
## 14) diabetes< 0.5 15717 8966 B1 (0.43 0.35 0.15 0.064 0.0071)
## 28) cancer< 0.5 13123 7034 B1 (0.46 0.34 0.13 0.058 0.0065)
## 56) arthritis< 0.5 9625 4692 B1 (0.51 0.31 0.12 0.054 0.0058)
## 112) ihd< 0.5 3135 1246 B1 (0.6 0.26 0.095 0.036 0.0032)
## 224) depression< 0.5 2292 821 B1 (0.64 0.24 0.08 0.034 0.0044) *
## 225) depression>=0.5 843 425 B1 (0.5 0.33 0.14 0.04 0)
## 450) age< 92.5 810 398 B1 (0.51 0.32 0.13 0.041 0)
## 900) reimbursement2008>=11525 117 40 B1 (0.66 0.21 0.068 0.06 0) *
## 901) reimbursement2008< 11525 693 358 B1 (0.48 0.33 0.14 0.038 0)
## 1802) reimbursement2008< 11105 684 352 B1 (0.49 0.34 0.14 0.038 0)
## 3604) reimbursement2008< 4365 286 134 B1 (0.53 0.33 0.12 0.017 0) *
## 3605) reimbursement2008>=4365 398 218 B1 (0.45 0.34 0.15 0.053 0)
## 7210) reimbursement2008>=4700 340 173 B1 (0.49 0.31 0.14 0.053 0) *
## 7211) reimbursement2008< 4700 58 28 B2 (0.22 0.52 0.21 0.052 0) *
## 1803) reimbursement2008>=11105 9 4 B3 (0.33 0.11 0.56 0 0) *
## 451) age>=92.5 33 14 B2 (0.18 0.58 0.21 0.03 0) *
## 113) ihd>=0.5 6490 3446 B1 (0.47 0.33 0.13 0.063 0.0071)
## 226) depression< 0.5 4266 2110 B1 (0.51 0.31 0.12 0.056 0.0061)
## 452) osteoporosis< 0.5 3304 1572 B1 (0.52 0.3 0.12 0.055 0.0064)
## 904) reimbursement2008>=5905 1626 714 B1 (0.56 0.25 0.12 0.061 0.0068) *
## 905) reimbursement2008< 5905 1678 858 B1 (0.49 0.34 0.12 0.05 0.006)
## 1810) reimbursement2008< 5695 1608 814 B1 (0.49 0.33 0.12 0.051 0.0062) *
## 1811) reimbursement2008>=5695 70 34 B2 (0.37 0.51 0.086 0.029 0) *
## 453) osteoporosis>=0.5 962 538 B1 (0.44 0.38 0.12 0.057 0.0052)
## 906) stroke< 0.5 857 465 B1 (0.46 0.37 0.12 0.056 0.0047)
## 1812) heart.failure< 0.5 405 203 B1 (0.5 0.35 0.1 0.044 0.0074)
## 3624) age< 83.5 329 159 B1 (0.52 0.32 0.11 0.049 0.0091) *
## 3625) age>=83.5 76 41 B2 (0.42 0.46 0.092 0.026 0)
## 7250) reimbursement2008>=6785 21 7 B1 (0.67 0.24 0.095 0 0) *
## 7251) reimbursement2008< 6785 55 25 B2 (0.33 0.55 0.091 0.036 0) *
## 1813) heart.failure>=0.5 452 262 B1 (0.42 0.38 0.13 0.066 0.0022)
## 3626) reimbursement2008>=3875 362 201 B1 (0.44 0.35 0.13 0.069 0.0028) *
## 3627) reimbursement2008< 3875 90 45 B2 (0.32 0.5 0.12 0.056 0)
## 7254) age< 69.5 21 9 B1 (0.57 0.29 0.048 0.095 0) *
## 7255) age>=69.5 69 30 B2 (0.25 0.57 0.14 0.043 0) *
## 907) stroke>=0.5 105 54 B2 (0.3 0.49 0.13 0.067 0.0095) *
## 227) depression>=0.5 2224 1336 B1 (0.4 0.35 0.16 0.076 0.009)
## 454) kidney< 0.5 1518 863 B1 (0.43 0.34 0.16 0.061 0.0053) *
## 455) kidney>=0.5 706 440 B2 (0.33 0.38 0.17 0.11 0.017)
## 910) reimbursement2008>=3155 696 431 B2 (0.33 0.38 0.16 0.11 0.017)
## 1820) heart.failure< 0.5 177 99 B1 (0.44 0.35 0.15 0.062 0) *
## 1821) heart.failure>=0.5 519 316 B2 (0.3 0.39 0.17 0.12 0.023) *
## 911) reimbursement2008< 3155 10 4 B3 (0.1 0.1 0.6 0.2 0) *
## 57) arthritis>=0.5 3498 2017 B2 (0.33 0.42 0.17 0.069 0.0083)
## 114) reimbursement2008< 8525 2340 1270 B2 (0.31 0.46 0.17 0.062 0.0064)
## 228) reimbursement2008< 4645 1359 754 B2 (0.34 0.45 0.15 0.056 0.0059)
## 456) ihd< 0.5 440 248 B2 (0.4 0.44 0.11 0.045 0.0045)
## 912) reimbursement2008< 3155 58 22 B2 (0.34 0.62 0 0.017 0.017) *
## 913) reimbursement2008>=3155 382 225 B1 (0.41 0.41 0.13 0.05 0.0026)
## 1826) reimbursement2008< 3245 25 8 B1 (0.68 0.28 0.04 0 0) *
## 1827) reimbursement2008>=3245 357 208 B2 (0.39 0.42 0.13 0.053 0.0028)
## 3654) age>=80.5 91 46 B1 (0.49 0.34 0.099 0.066 0) *
## 3655) age< 80.5 266 148 B2 (0.36 0.44 0.15 0.049 0.0038) *
## 457) ihd>=0.5 919 506 B2 (0.32 0.45 0.17 0.061 0.0065) *
## 229) reimbursement2008>=4645 981 516 B2 (0.26 0.47 0.19 0.069 0.0071) *
## 115) reimbursement2008>=8525 1158 722 B1 (0.38 0.35 0.17 0.085 0.012)
## 230) copd< 0.5 714 396 B1 (0.45 0.33 0.13 0.085 0.007)
## 460) depression< 0.5 412 196 B1 (0.52 0.29 0.1 0.073 0.0097) *
## 461) depression>=0.5 302 183 B2 (0.34 0.39 0.16 0.1 0.0033)
## 922) age>=92.5 9 3 B1 (0.67 0 0.11 0.11 0.11) *
## 923) age< 92.5 293 174 B2 (0.33 0.41 0.16 0.1 0)
## 1846) stroke>=0.5 39 19 B1 (0.51 0.31 0.1 0.077 0) *
## 1847) stroke< 0.5 254 147 B2 (0.3 0.42 0.17 0.11 0) *
## 231) copd>=0.5 444 272 B2 (0.27 0.39 0.24 0.086 0.02)
## 462) osteoporosis< 0.5 282 187 B2 (0.31 0.34 0.25 0.082 0.018)
## 924) reimbursement2008< 27390 220 143 B1 (0.35 0.3 0.26 0.073 0.018)
## 1848) reimbursement2008>=12810 132 78 B1 (0.41 0.32 0.2 0.068 0.0076)
## 3696) age< 84.5 105 55 B1 (0.48 0.33 0.15 0.029 0.0095) *
## 3697) age>=84.5 27 17 B3 (0.15 0.26 0.37 0.22 0) *
## 1849) reimbursement2008< 12810 88 57 B3 (0.26 0.27 0.35 0.08 0.034) *
## 925) reimbursement2008>=27390 62 33 B2 (0.18 0.47 0.23 0.11 0.016) *
## 463) osteoporosis>=0.5 162 85 B2 (0.19 0.48 0.22 0.093 0.025) *
## 29) cancer>=0.5 2594 1539 B2 (0.26 0.41 0.24 0.091 0.01)
## 58) reimbursement2008< 5770 1000 562 B2 (0.3 0.44 0.19 0.07 0.005) *
## 59) reimbursement2008>=5770 1594 977 B2 (0.23 0.39 0.27 0.1 0.014)
## 118) reimbursement2008>=8645 1054 656 B2 (0.27 0.38 0.24 0.1 0.015)
## 236) arthritis< 0.5 745 464 B2 (0.31 0.38 0.2 0.097 0.013)
## 472) ihd< 0.5 159 94 B1 (0.41 0.32 0.21 0.05 0.013)
## 944) reimbursement2008>=11995 76 36 B1 (0.53 0.24 0.16 0.066 0.013) *
## 945) reimbursement2008< 11995 83 50 B2 (0.3 0.4 0.25 0.036 0.012) *
## 473) ihd>=0.5 586 356 B2 (0.28 0.39 0.2 0.11 0.014) *
## 237) arthritis>=0.5 309 192 B2 (0.16 0.38 0.32 0.12 0.019)
## 474) reimbursement2008>=10960 237 136 B2 (0.15 0.43 0.3 0.11 0.013)
## 948) copd< 0.5 126 64 B2 (0.17 0.49 0.26 0.071 0) *
## 949) copd>=0.5 111 72 B2 (0.12 0.35 0.35 0.15 0.027)
## 1898) age< 75.5 54 30 B3 (0.15 0.26 0.44 0.13 0.019) *
## 1899) age>=75.5 57 32 B2 (0.088 0.44 0.26 0.18 0.035) *
## 475) reimbursement2008< 10960 72 44 B3 (0.19 0.22 0.39 0.15 0.042) *
## 119) reimbursement2008< 8645 540 321 B2 (0.16 0.41 0.32 0.11 0.011)
## 238) heart.failure>=0.5 243 128 B2 (0.14 0.47 0.28 0.099 0.016) *
## 239) heart.failure< 0.5 297 191 B3 (0.18 0.35 0.36 0.11 0.0067)
## 478) depression< 0.5 226 141 B2 (0.18 0.38 0.33 0.12 0.0044) *
## 479) depression>=0.5 71 39 B3 (0.17 0.27 0.45 0.099 0.014) *
## 15) diabetes>=0.5 53801 31450 B2 (0.23 0.42 0.21 0.13 0.02)
## 30) kidney< 0.5 25067 14311 B2 (0.3 0.43 0.19 0.076 0.0074)
## 60) arthritis< 0.5 15178 9179 B2 (0.35 0.4 0.17 0.069 0.0063)
## 120) cancer< 0.5 12572 7709 B2 (0.39 0.39 0.16 0.063 0.0059)
## 240) ihd< 0.5 2617 1376 B1 (0.47 0.34 0.13 0.049 0.0053)
## 480) reimbursement2008>=9400 403 171 B1 (0.58 0.21 0.15 0.05 0.0099) *
## 481) reimbursement2008< 9400 2214 1205 B1 (0.46 0.36 0.13 0.048 0.0045)
## 962) osteoporosis< 0.5 1636 847 B1 (0.48 0.34 0.12 0.049 0.0043)
## 1924) alzheimers< 0.5 1127 559 B1 (0.5 0.33 0.12 0.042 0.0035) *
## 1925) alzheimers>=0.5 509 288 B1 (0.43 0.36 0.13 0.065 0.0059)
## 3850) reimbursement2008< 3775 137 68 B1 (0.5 0.3 0.12 0.066 0.0073) *
## 3851) reimbursement2008>=3775 372 220 B1 (0.41 0.39 0.13 0.065 0.0054)
## 7702) reimbursement2008>=4055 330 188 B1 (0.43 0.36 0.13 0.07 0.0061)
## 15404) reimbursement2008>=4185 309 177 B1 (0.43 0.38 0.12 0.065 0.0065)
## 30808) reimbursement2008>=4635 253 138 B1 (0.45 0.35 0.12 0.067 0.0079)
## 61616) age< 96 245 131 B1 (0.47 0.36 0.11 0.065 0.0041)
## 123232) reimbursement2008< 8170 209 106 B1 (0.49 0.33 0.11 0.067 0) *
## 123233) reimbursement2008>=8170 36 19 B2 (0.31 0.47 0.14 0.056 0.028)
## 246466) age< 74.5 15 6 B1 (0.6 0.13 0.2 0 0.067) *
## 246467) age>=74.5 21 6 B2 (0.095 0.71 0.095 0.095 0) *
## 61617) age>=96 8 5 B3 (0.12 0.25 0.38 0.12 0.12) *
## 30809) reimbursement2008< 4635 56 28 B2 (0.3 0.5 0.14 0.054 0) *
## 15405) reimbursement2008< 4185 21 11 B1 (0.48 0.095 0.29 0.14 0) *
## 7703) reimbursement2008< 4055 42 17 B2 (0.24 0.6 0.14 0.024 0) *
## 963) osteoporosis>=0.5 578 342 B2 (0.38 0.41 0.16 0.047 0.0052)
## 1926) depression< 0.5 339 189 B1 (0.44 0.37 0.13 0.047 0.0029)
## 3852) reimbursement2008< 4905 211 119 B1 (0.44 0.42 0.11 0.033 0)
## 7704) reimbursement2008< 4075 142 71 B1 (0.5 0.36 0.11 0.035 0) *
## 7705) reimbursement2008>=4075 69 31 B2 (0.3 0.55 0.12 0.029 0) *
## 3853) reimbursement2008>=4905 128 70 B1 (0.45 0.3 0.17 0.07 0.0078) *
## 1927) depression>=0.5 239 130 B2 (0.29 0.46 0.2 0.046 0.0084)
## 3854) copd< 0.5 181 88 B2 (0.31 0.51 0.13 0.039 0.011) *
## 3855) copd>=0.5 58 34 B3 (0.24 0.28 0.41 0.069 0) *
## 241) ihd>=0.5 9955 5976 B2 (0.36 0.4 0.17 0.067 0.006)
## 482) depression< 0.5 5563 3339 B1 (0.4 0.38 0.15 0.059 0.0059)
## 964) reimbursement2008>=8955 1363 758 B1 (0.44 0.32 0.16 0.067 0.0088)
## 1928) copd< 0.5 798 405 B1 (0.49 0.32 0.12 0.064 0.0038) *
## 1929) copd>=0.5 565 353 B1 (0.38 0.32 0.22 0.073 0.016)
## 3858) stroke>=0.5 116 64 B2 (0.35 0.45 0.12 0.06 0.017)
## 7716) age>=74.5 63 36 B1 (0.43 0.33 0.16 0.063 0.016) *
## 7717) age< 74.5 53 22 B2 (0.26 0.58 0.075 0.057 0.019) *
## 3859) stroke< 0.5 449 278 B1 (0.38 0.28 0.24 0.076 0.016) *
## 965) reimbursement2008< 8955 4200 2510 B2 (0.39 0.4 0.15 0.056 0.005)
## 1930) heart.failure< 0.5 1953 1129 B1 (0.42 0.4 0.13 0.045 0.0041)
## 3860) reimbursement2008< 3415 343 172 B1 (0.5 0.37 0.096 0.032 0.0058) *
## 3861) reimbursement2008>=3415 1610 954 B2 (0.41 0.41 0.14 0.048 0.0037)
## 7722) age< 42.5 43 17 B1 (0.6 0.26 0.07 0.07 0) *
## 7723) age>=42.5 1567 922 B2 (0.4 0.41 0.14 0.047 0.0038)
## 15446) age>=50.5 1527 894 B2 (0.4 0.41 0.13 0.047 0.0039)
## 30892) reimbursement2008>=3465 1478 870 B2 (0.41 0.41 0.13 0.045 0.0027)
## 61784) reimbursement2008< 4655 759 431 B1 (0.43 0.4 0.12 0.043 0.0013)
## 123568) reimbursement2008>=4315 158 79 B1 (0.5 0.35 0.095 0.051 0.0063) *
## 123569) reimbursement2008< 4315 601 352 B1 (0.41 0.41 0.13 0.042 0)
## 247138) reimbursement2008< 4295 592 344 B1 (0.42 0.41 0.13 0.042 0)
## 494276) age>=82.5 135 71 B1 (0.47 0.36 0.15 0.015 0) *
## 494277) age< 82.5 457 266 B2 (0.4 0.42 0.13 0.05 0)
## 988554) age< 74.5 290 166 B1 (0.43 0.39 0.13 0.052 0)
## 1977108) age>=62.5 234 128 B1 (0.45 0.38 0.12 0.047 0) *
## 1977109) age< 62.5 56 31 B2 (0.32 0.45 0.16 0.071 0) *
## 988555) age>=74.5 167 90 B2 (0.36 0.46 0.13 0.048 0)
## 1977110) reimbursement2008>=4105 39 20 B1 (0.49 0.36 0.13 0.026 0) *
## 1977111) reimbursement2008< 4105 128 65 B2 (0.32 0.49 0.13 0.055 0) *
## 247139) reimbursement2008>=4295 9 1 B2 (0.11 0.89 0 0 0) *
## 61785) reimbursement2008>=4655 719 414 B2 (0.38 0.42 0.15 0.047 0.0042)
## 123570) reimbursement2008< 5835 346 180 B2 (0.35 0.48 0.13 0.038 0.0029) *
## 123571) reimbursement2008>=5835 373 223 B1 (0.4 0.37 0.16 0.056 0.0054)
## 247142) alzheimers>=0.5 124 64 B1 (0.48 0.31 0.15 0.04 0.0081)
## 494284) reimbursement2008< 8555 114 55 B1 (0.52 0.28 0.16 0.035 0.0088) *
## 494285) reimbursement2008>=8555 10 3 B2 (0.1 0.7 0.1 0.1 0) *
## 247143) alzheimers< 0.5 249 149 B2 (0.36 0.4 0.17 0.064 0.004)
## 494286) reimbursement2008>=6045 217 124 B2 (0.36 0.43 0.14 0.069 0.0046) *
## 494287) reimbursement2008< 6045 32 20 B1 (0.38 0.22 0.38 0.031 0)
## 988574) age< 72.5 11 4 B1 (0.64 0.18 0.18 0 0) *
## 988575) age>=72.5 21 11 B3 (0.24 0.24 0.48 0.048 0) *
## 30893) reimbursement2008< 3465 49 24 B2 (0.27 0.51 0.082 0.1 0.041) *
## 15447) age< 50.5 40 26 B1 (0.35 0.3 0.3 0.05 0) *
## 1931) heart.failure>=0.5 2247 1339 B2 (0.35 0.4 0.17 0.066 0.0058)
## 3862) reimbursement2008>=5335 866 530 B1 (0.39 0.37 0.16 0.074 0.0058)
## 7724) reimbursement2008>=8115 129 68 B2 (0.36 0.47 0.12 0.047 0) *
## 7725) reimbursement2008< 8115 737 447 B1 (0.39 0.35 0.17 0.079 0.0068)
## 15450) age< 94.5 703 421 B1 (0.4 0.35 0.17 0.075 0.0071)
## 30900) reimbursement2008>=6635 298 164 B1 (0.45 0.32 0.15 0.067 0.013) *
## 30901) reimbursement2008< 6635 405 255 B2 (0.37 0.37 0.18 0.081 0.0025)
## 61802) reimbursement2008< 5685 137 80 B1 (0.42 0.31 0.16 0.11 0) *
## 61803) reimbursement2008>=5685 268 161 B2 (0.34 0.4 0.19 0.067 0.0037) *
## 15451) age>=94.5 34 17 B2 (0.24 0.5 0.12 0.15 0) *
## 3863) reimbursement2008< 5335 1381 795 B2 (0.33 0.42 0.18 0.061 0.0058)
## 7726) copd< 0.5 997 591 B2 (0.36 0.41 0.17 0.057 0.006)
## 15452) age< 69.5 297 171 B1 (0.42 0.38 0.15 0.04 0.0034)
## 30904) reimbursement2008< 5065 274 153 B1 (0.44 0.36 0.15 0.04 0.0036)
## 61808) alzheimers< 0.5 174 91 B1 (0.48 0.3 0.17 0.046 0.0057) *
## 61809) alzheimers>=0.5 100 54 B2 (0.38 0.46 0.13 0.03 0)
## 123618) reimbursement2008>=4355 26 10 B1 (0.62 0.15 0.15 0.077 0) *
## 123619) reimbursement2008< 4355 74 32 B2 (0.3 0.57 0.12 0.014 0) *
## 30905) reimbursement2008>=5065 23 9 B2 (0.22 0.61 0.13 0.043 0) *
## 15453) age>=69.5 700 407 B2 (0.33 0.42 0.18 0.064 0.0071) *
## 7727) copd>=0.5 384 204 B2 (0.27 0.47 0.19 0.07 0.0052) *
## 483) depression>=0.5 4392 2538 B2 (0.31 0.42 0.18 0.077 0.0061)
## 966) reimbursement2008< 8325 2928 1619 B2 (0.31 0.45 0.17 0.065 0.0065)
## 1932) copd< 0.5 1987 1102 B2 (0.33 0.45 0.16 0.056 0.0045)
## 3864) age< 98.5 1964 1085 B2 (0.33 0.45 0.16 0.057 0.0046)
## 7728) reimbursement2008< 3085 22 8 B1 (0.64 0.23 0.045 0.091 0) *
## 7729) reimbursement2008>=3085 1942 1068 B2 (0.33 0.45 0.16 0.056 0.0046)
## 15458) heart.failure< 0.5 889 508 B2 (0.37 0.43 0.15 0.051 0.0034) *
## 15459) heart.failure>=0.5 1053 560 B2 (0.3 0.47 0.17 0.061 0.0057)
## 30918) osteoporosis< 0.5 721 396 B2 (0.32 0.45 0.16 0.061 0.0055)
## 61836) age>=86.5 109 59 B1 (0.46 0.3 0.15 0.083 0.0092) *
## 61837) age< 86.5 612 320 B2 (0.3 0.48 0.16 0.057 0.0049) *
## 30919) osteoporosis>=0.5 332 164 B2 (0.24 0.51 0.18 0.06 0.006) *
## 3865) age>=98.5 23 12 B3 (0.26 0.26 0.48 0 0) *
## 1933) copd>=0.5 941 517 B2 (0.26 0.45 0.2 0.084 0.011) *
## 967) reimbursement2008>=8325 1464 919 B2 (0.32 0.37 0.2 0.1 0.0055)
## 1934) reimbursement2008< 8485 36 16 B1 (0.56 0.22 0.22 0 0) *
## 1935) reimbursement2008>=8485 1428 891 B2 (0.32 0.38 0.2 0.1 0.0056)
## 3870) age< 78.5 837 532 B2 (0.35 0.36 0.19 0.098 0.0036)
## 7740) reimbursement2008< 21320 639 406 B1 (0.36 0.35 0.19 0.092 0.0031)
## 15480) age< 49.5 83 47 B2 (0.35 0.43 0.096 0.12 0) *
## 15481) age>=49.5 556 352 B1 (0.37 0.33 0.21 0.088 0.0036)
## 30962) age>=67.5 368 230 B1 (0.38 0.36 0.18 0.087 0)
## 61924) reimbursement2008>=10440 261 153 B1 (0.41 0.33 0.17 0.084 0)
## 123848) reimbursement2008< 12585 92 46 B1 (0.5 0.25 0.18 0.065 0) *
## 123849) reimbursement2008>=12585 169 105 B2 (0.37 0.38 0.16 0.095 0)
## 247698) reimbursement2008>=14485 109 63 B1 (0.42 0.31 0.17 0.092 0)
## 495396) osteoporosis< 0.5 72 37 B1 (0.49 0.24 0.17 0.11 0) *
## 495397) osteoporosis>=0.5 37 20 B2 (0.3 0.46 0.19 0.054 0) *
## 247699) reimbursement2008< 14485 60 30 B2 (0.27 0.5 0.13 0.1 0) *
## 61925) reimbursement2008< 10440 107 62 B2 (0.28 0.42 0.21 0.093 0) *
## 30963) age< 67.5 188 122 B1 (0.35 0.28 0.27 0.09 0.011)
## 61926) age>=55.5 135 82 B1 (0.39 0.25 0.27 0.089 0) *
## 61927) age< 55.5 53 34 B2 (0.25 0.36 0.26 0.094 0.038) *
## 7741) reimbursement2008>=21320 198 114 B2 (0.3 0.42 0.16 0.12 0.0051) *
## 3871) age>=78.5 591 359 B2 (0.28 0.39 0.21 0.11 0.0085)
## 7742) heart.failure< 0.5 122 73 B1 (0.4 0.32 0.18 0.098 0)
## 15484) reimbursement2008< 11560 40 17 B1 (0.58 0.2 0.18 0.05 0) *
## 15485) reimbursement2008>=11560 82 51 B2 (0.32 0.38 0.18 0.12 0) *
## 7743) heart.failure>=0.5 469 276 B2 (0.24 0.41 0.22 0.11 0.011) *
## 121) cancer>=0.5 2606 1470 B2 (0.21 0.44 0.25 0.098 0.0081) *
## 61) arthritis>=0.5 9889 5132 B2 (0.22 0.48 0.21 0.088 0.0092)
## 122) depression< 0.5 5134 2665 B2 (0.25 0.48 0.18 0.08 0.0078)
## 244) cancer< 0.5 4305 2260 B2 (0.27 0.48 0.17 0.076 0.0086)
## 488) reimbursement2008>=9880 1063 636 B2 (0.32 0.4 0.18 0.089 0.012)
## 976) ihd< 0.5 102 49 B1 (0.52 0.27 0.13 0.069 0.0098) *
## 977) ihd>=0.5 961 562 B2 (0.29 0.42 0.19 0.092 0.012) *
## 489) reimbursement2008< 9880 3242 1624 B2 (0.25 0.5 0.17 0.072 0.0074) *
## 245) cancer>=0.5 829 405 B2 (0.15 0.51 0.23 0.1 0.0036) *
## 123) depression>=0.5 4755 2467 B2 (0.18 0.48 0.23 0.096 0.011) *
## 31) kidney>=0.5 28734 17139 B2 (0.16 0.4 0.23 0.17 0.03)
## 62) reimbursement2008< 15395 16249 9131 B2 (0.19 0.44 0.24 0.12 0.016)
## 124) arthritis< 0.5 9424 5647 B2 (0.23 0.4 0.23 0.12 0.017)
## 248) cancer< 0.5 7786 4711 B2 (0.25 0.39 0.21 0.12 0.017)
## 496) ihd< 0.5 964 608 B1 (0.37 0.36 0.17 0.085 0.011)
## 992) depression< 0.5 572 338 B1 (0.41 0.32 0.16 0.1 0.01)
## 1984) reimbursement2008< 3545 101 53 B2 (0.33 0.48 0.15 0.04 0.0099) *
## 1985) reimbursement2008>=3545 471 270 B1 (0.43 0.29 0.16 0.11 0.011)
## 3970) osteoporosis< 0.5 346 186 B1 (0.46 0.27 0.15 0.11 0.014) *
## 3971) osteoporosis>=0.5 125 82 B2 (0.33 0.34 0.2 0.13 0)
## 7942) age>=62 106 67 B1 (0.37 0.37 0.15 0.11 0)
## 15884) age>=67.5 93 55 B2 (0.34 0.41 0.16 0.086 0)
## 31768) reimbursement2008>=6110 44 23 B1 (0.48 0.3 0.11 0.11 0)
## 63536) reimbursement2008< 9180 26 9 B1 (0.65 0.15 0.077 0.12 0) *
## 63537) reimbursement2008>=9180 18 9 B2 (0.22 0.5 0.17 0.11 0) *
## 31769) reimbursement2008< 6110 49 24 B2 (0.22 0.51 0.2 0.061 0) *
## 15885) age< 67.5 13 6 B1 (0.54 0.077 0.077 0.31 0) *
## 7943) age< 62 19 10 B3 (0.11 0.21 0.47 0.21 0) *
## 993) depression>=0.5 392 227 B2 (0.31 0.42 0.19 0.064 0.013)
## 1986) reimbursement2008>=14460 9 2 B1 (0.78 0.22 0 0 0) *
## 1987) reimbursement2008< 14460 383 220 B2 (0.3 0.43 0.2 0.065 0.013) *
## 497) ihd>=0.5 6822 4095 B2 (0.24 0.4 0.22 0.12 0.018)
## 994) reimbursement2008< 6325 3172 1786 B2 (0.22 0.44 0.22 0.11 0.016) *
## 995) reimbursement2008>=6325 3650 2309 B2 (0.25 0.37 0.22 0.14 0.019)
## 1990) osteoporosis< 0.5 2424 1594 B2 (0.27 0.34 0.23 0.14 0.02)
## 3980) depression< 0.5 1234 816 B2 (0.3 0.34 0.2 0.13 0.024)
## 7960) reimbursement2008>=12135 349 226 B1 (0.35 0.3 0.16 0.16 0.032)
## 15920) age>=54 331 210 B1 (0.37 0.28 0.16 0.16 0.03) *
## 15921) age< 54 18 9 B2 (0.11 0.5 0.11 0.22 0.056) *
## 7961) reimbursement2008< 12135 885 570 B2 (0.28 0.36 0.22 0.12 0.021) *
## 3981) depression>=0.5 1190 778 B2 (0.24 0.35 0.25 0.15 0.016)
## 7962) copd< 0.5 547 367 B2 (0.28 0.33 0.25 0.12 0.022)
## 15924) reimbursement2008>=9205 310 209 B1 (0.33 0.32 0.21 0.12 0.029)
## 31848) reimbursement2008< 9955 50 28 B2 (0.42 0.44 0.02 0.12 0) *
## 31849) reimbursement2008>=9955 260 180 B1 (0.31 0.3 0.24 0.12 0.035)
## 63698) reimbursement2008>=14765 20 9 B1 (0.55 0.25 0.1 0.05 0.05) *
## 63699) reimbursement2008< 14765 240 168 B2 (0.29 0.3 0.25 0.12 0.033)
## 127398) age>=61.5 201 138 B1 (0.31 0.29 0.24 0.13 0.02)
## 254796) reimbursement2008< 12625 112 69 B1 (0.38 0.24 0.28 0.089 0.0089) *
## 254797) reimbursement2008>=12625 89 58 B2 (0.22 0.35 0.2 0.19 0.034) *
## 127399) age< 61.5 39 25 B2 (0.15 0.36 0.31 0.077 0.1) *
## 15925) reimbursement2008< 9205 237 156 B2 (0.22 0.34 0.3 0.12 0.013)
## 31850) age< 67.5 56 29 B2 (0.2 0.48 0.16 0.16 0) *
## 31851) age>=67.5 181 118 B3 (0.23 0.3 0.35 0.1 0.017)
## 63702) reimbursement2008>=6865 136 82 B3 (0.25 0.26 0.4 0.074 0.015) *
## 63703) reimbursement2008< 6865 45 27 B2 (0.18 0.4 0.2 0.2 0.022) *
## 7963) copd>=0.5 643 411 B2 (0.21 0.36 0.25 0.17 0.011) *
## 1991) osteoporosis>=0.5 1226 715 B2 (0.21 0.42 0.22 0.14 0.017) *
## 249) cancer>=0.5 1638 936 B2 (0.13 0.43 0.29 0.14 0.016) *
## 125) arthritis>=0.5 6825 3484 B2 (0.13 0.49 0.25 0.12 0.014) *
## 63) reimbursement2008>=15395 12485 8008 B2 (0.13 0.36 0.23 0.24 0.049)
## 126) arthritis>=0.5 5402 3220 B2 (0.094 0.4 0.24 0.22 0.04)
## 252) reimbursement2008< 34925 3345 1942 B2 (0.11 0.42 0.25 0.19 0.03)
## 504) depression< 0.5 1291 714 B2 (0.14 0.45 0.22 0.17 0.025)
## 1008) cancer< 0.5 973 546 B2 (0.16 0.44 0.19 0.18 0.029) *
## 1009) cancer>=0.5 318 168 B2 (0.072 0.47 0.3 0.14 0.013)
## 2018) reimbursement2008>=16525 293 144 B2 (0.068 0.51 0.28 0.13 0.014) *
## 2019) reimbursement2008< 16525 25 10 B3 (0.12 0.04 0.6 0.24 0) *
## 505) depression>=0.5 2054 1228 B2 (0.092 0.4 0.27 0.2 0.034) *
## 253) reimbursement2008>=34925 2057 1278 B2 (0.067 0.38 0.23 0.27 0.055)
## 506) copd< 0.5 520 300 B2 (0.096 0.42 0.23 0.21 0.042) *
## 507) copd>=0.5 1537 978 B2 (0.057 0.36 0.23 0.29 0.06)
## 1014) age>=62.5 1286 804 B2 (0.058 0.37 0.24 0.27 0.061) *
## 1015) age< 62.5 251 153 B4 (0.052 0.31 0.2 0.39 0.052)
## 2030) reimbursement2008< 101585 237 150 B4 (0.055 0.32 0.21 0.37 0.051)
## 4060) cancer>=0.5 62 36 B2 (0.048 0.42 0.27 0.24 0.016) *
## 4061) cancer< 0.5 175 103 B4 (0.057 0.29 0.18 0.41 0.063) *
## 2031) reimbursement2008>=101585 14 3 B4 (0 0.071 0.071 0.79 0.071) *
## 127) arthritis< 0.5 7083 4788 B2 (0.15 0.32 0.22 0.25 0.057)
## 254) cancer< 0.5 5298 3651 B2 (0.17 0.31 0.2 0.26 0.062)
## 508) depression< 0.5 2489 1797 B2 (0.22 0.28 0.18 0.27 0.06)
## 1016) copd>=0.5 1317 890 B2 (0.2 0.32 0.18 0.24 0.056)
## 2032) ihd< 0.5 72 41 B1 (0.43 0.25 0.15 0.11 0.056) *
## 2033) ihd>=0.5 1245 836 B2 (0.19 0.33 0.18 0.24 0.056) *
## 1017) copd< 0.5 1172 815 B4 (0.23 0.23 0.17 0.3 0.065)
## 2034) reimbursement2008>=43640 191 129 B2 (0.15 0.32 0.23 0.22 0.073)
## 4068) age>=64.5 172 112 B2 (0.16 0.35 0.2 0.23 0.064) *
## 4069) age< 64.5 19 10 B3 (0.11 0.11 0.47 0.16 0.16) *
## 2035) reimbursement2008< 43640 981 666 B4 (0.25 0.21 0.16 0.32 0.063)
## 4070) reimbursement2008< 23175 468 337 B1 (0.28 0.24 0.16 0.26 0.056)
## 8140) age< 93.5 457 326 B1 (0.29 0.23 0.16 0.26 0.057)
## 16280) age< 86.5 398 288 B1 (0.28 0.25 0.16 0.25 0.063)
## 32560) alzheimers>=0.5 179 122 B1 (0.32 0.28 0.15 0.17 0.073)
## 65120) reimbursement2008>=21440 38 19 B2 (0.24 0.5 0.11 0.13 0.026) *
## 65121) reimbursement2008< 21440 141 93 B1 (0.34 0.23 0.16 0.18 0.085)
## 130242) reimbursement2008>=17585 89 53 B1 (0.4 0.17 0.19 0.16 0.079) *
## 130243) reimbursement2008< 17585 52 35 B2 (0.23 0.33 0.12 0.23 0.096) *
## 32561) alzheimers< 0.5 219 151 B4 (0.24 0.23 0.16 0.31 0.055) *
## 16281) age>=86.5 59 38 B1 (0.36 0.1 0.17 0.36 0.017)
## 32562) reimbursement2008>=19680 18 8 B1 (0.56 0.11 0 0.28 0.056) *
## 32563) reimbursement2008< 19680 41 25 B4 (0.27 0.098 0.24 0.39 0) *
## 8141) age>=93.5 11 6 B2 (0 0.45 0.36 0.18 0) *
## 4071) reimbursement2008>=23175 513 320 B4 (0.22 0.18 0.16 0.38 0.07) *
## 509) depression>=0.5 2809 1854 B2 (0.13 0.34 0.22 0.25 0.063) *
## 255) cancer>=0.5 1785 1137 B2 (0.097 0.36 0.28 0.22 0.041) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 171981 12159 140 186 0
## B2 26534 25373 156 196 0
## B3 12021 12119 286 160 0
## B4 4652 6824 70 360 0
## B5 483 1029 14 60 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7205162 NA 0.7188342 0.7221934 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 0.0000000## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114141 8610 124 103 0
## B2 18409 16102 187 142 0
## B3 8027 8146 118 99 0
## B4 3099 4584 53 201 0
## B5 351 657 4 45 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.126669e-01 NA 7.105887e-01 7.147384e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 7.015238e-319 0.000000e+00
## model_id model_method
## 1 All.X.lser.no.cp.4015.rpart rpart
## feats
## 1 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 1 73.967 15.63
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.7117062 0.7188342 0.7221934
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 0.3328722 0.7126669 0.7105887
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.7147384 NA 0 0.002149621
## max.KappaSD.fit
## 1 0.00733801
## [1] "fitting model: All.X.lser.ys.cp.opt.rpart"
## [1] " indep_vars: age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008"
## + Fold1: cp=0.00435
## - Fold1: cp=0.00435
## + Fold2: cp=0.00435
## - Fold2: cp=0.00435
## + Fold3: cp=0.00435
## - Fold3: cp=0.00435
## + Fold4: cp=0.00435
## - Fold4: cp=0.00435
## + Fold5: cp=0.00435
## - Fold5: cp=0.00435
## Aggregating results
## Selecting tuning parameters
## Fitting cp = 0.014 on full training set
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 0.04908841 0 1.0000000
## 2 0.01395884 2 0.9018232
##
## Variable importance
## reimbursement2008 bucket2008 ihd diabetes
## 31 21 14 13
## heart.failure kidney
## 11 9
##
## Node number 1: 274803 observations, complexity param=0.04908841
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (165987 obs) right son=3 (108816 obs)
## Primary splits:
## reimbursement2008 < 1565 to the left, improve=24395.14, (0 missing)
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## ihd < 0.5 to the left, improve=16291.74, (0 missing)
## diabetes < 0.5 to the left, improve=16041.26, (0 missing)
## heart.failure < 0.5 to the left, improve=12498.16, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.861, adj=0.650, (0 split)
## ihd < 0.5 to the left, agree=0.792, adj=0.474, (0 split)
## diabetes < 0.5 to the left, agree=0.785, adj=0.456, (0 split)
## heart.failure < 0.5 to the left, agree=0.762, adj=0.399, (0 split)
## kidney < 0.5 to the left, agree=0.731, adj=0.321, (0 split)
##
## Node number 2: 165987 observations
## predicted class=B1 expected loss=0.1261424 P(node) =0.6040218
## class counts: 145049 12284 6102 2315 237
## probabilities: 0.874 0.074 0.037 0.014 0.001
##
## Node number 3: 108816 observations, complexity param=0.04908841
## predicted class=B2 expected loss=0.6326367 P(node) =0.3959782
## class counts: 39417 39975 18484 9591 1349
## probabilities: 0.362 0.367 0.170 0.088 0.012
## left son=6 (39298 obs) right son=7 (69518 obs)
## Primary splits:
## reimbursement2008 < 3065 to the left, improve=2010.3080, (0 missing)
## bucket2008 < 1.5 to the left, improve=1980.9770, (0 missing)
## kidney < 0.5 to the left, improve=1416.9220, (0 missing)
## diabetes < 0.5 to the left, improve=1236.1460, (0 missing)
## heart.failure < 0.5 to the left, improve= 976.9427, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.989, adj=0.969, (0 split)
## ihd < 0.5 to the left, agree=0.659, adj=0.056, (0 split)
## diabetes < 0.5 to the left, agree=0.641, adj=0.006, (0 split)
##
## Node number 6: 39298 observations
## predicted class=B1 expected loss=0.4797445 P(node) =0.1430043
## class counts: 20445 12134 4756 1782 181
## probabilities: 0.520 0.309 0.121 0.045 0.005
##
## Node number 7: 69518 observations
## predicted class=B2 expected loss=0.5995138 P(node) =0.2529739
## class counts: 18972 27841 13728 7809 1168
## probabilities: 0.273 0.400 0.197 0.112 0.017
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) reimbursement2008< 1565 165987 20938 B1 (0.87 0.074 0.037 0.014 0.0014) *
## 3) reimbursement2008>=1565 108816 68841 B2 (0.36 0.37 0.17 0.088 0.012)
## 6) reimbursement2008< 3065 39298 18853 B1 (0.52 0.31 0.12 0.045 0.0046) *
## 7) reimbursement2008>=3065 69518 41677 B2 (0.27 0.4 0.2 0.11 0.017) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 165494 18972 0 0 0
## B2 24418 27841 0 0 0
## B3 10858 13728 0 0 0
## B4 4097 7809 0 0 0
## B5 418 1168 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.035404e-01 NA 7.018289e-01 7.052475e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 2.207353e-289 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110452 12526 0 0 0
## B2 16322 18518 0 0 0
## B3 7105 9285 0 0 0
## B4 2740 5197 0 0 0
## B5 299 758 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.039770e-01 NA 7.018807e-01 7.060669e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 6.148674e-199 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## model_id model_method
## 1 All.X.lser.ys.cp.opt.rpart rpart
## feats
## 1 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 3 108.644 15.505
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.7035404 0.7018289 0.7052475
## max.Kappa.fit min.loss.error.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.7784778 0.703977 0.7018807
## max.AccuracyUpper.OOB max.Kappa.OOB min.loss.error.OOB min.SSE.fit
## 1 0.7060669 NA 0.7441403 0
## min.loss.errorSD.fit
## 1 0.01642528
## [1] "fitting model: All.X.lser.ys.cp.4015.rpart"
## [1] " indep_vars: age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008"
## + Fold1: cp=5e-05
## - Fold1: cp=5e-05
## + Fold2: cp=5e-05
## - Fold2: cp=5e-05
## + Fold3: cp=5e-05
## - Fold3: cp=5e-05
## + Fold4: cp=5e-05
## - Fold4: cp=5e-05
## + Fold5: cp=5e-05
## - Fold5: cp=5e-05
## Aggregating results
## Fitting final model on full training set## Warning: labs do not fit even at cex 0.15, there may be some overplotting## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 4.908841e-02 0 1.0000000
## 2 1.395884e-02 2 0.9018232
## 3 4.350377e-03 3 0.8878643
## 4 3.597640e-03 4 0.8835140
## 5 9.879673e-04 5 0.8799163
## 6 7.859460e-04 10 0.8748685
## 7 6.789392e-04 11 0.8740826
## 8 4.649258e-04 14 0.8720458
## 9 4.372516e-04 15 0.8715809
## 10 2.988809e-04 19 0.8698319
## 11 2.767415e-04 20 0.8695330
## 12 2.435326e-04 21 0.8692562
## 13 2.036818e-04 22 0.8690127
## 14 1.881842e-04 28 0.8677729
## 15 1.826494e-04 29 0.8675847
## 16 1.605101e-04 31 0.8672194
## 17 1.439056e-04 33 0.8668984
## 18 1.411382e-04 37 0.8663228
## 19 1.217663e-04 42 0.8655922
## 20 1.162314e-04 45 0.8652269
## 21 1.129105e-04 47 0.8649944
## 22 1.051618e-04 52 0.8644299
## 23 9.962695e-05 57 0.8638764
## 24 9.409212e-05 68 0.8627473
## 25 8.855729e-05 74 0.8621827
## 26 8.302246e-05 83 0.8613746
## 27 7.748763e-05 85 0.8612086
## 28 7.379774e-05 97 0.8602455
## 29 7.195280e-05 101 0.8599134
## 30 6.918538e-05 114 0.8588729
## 31 6.641797e-05 122 0.8583194
## 32 6.272808e-05 154 0.8560612
## 33 6.167383e-05 158 0.8557955
## 34 6.088314e-05 166 0.8552752
## 35 5.811572e-05 179 0.8544340
## 36 5.534831e-05 183 0.8542015
## 37 5.313437e-05 228 0.8516001
## 38 5.258089e-05 233 0.8513344
## 39 5.165842e-05 237 0.8511241
## 40 5.000000e-05 254 0.8501832
##
## Variable importance
## reimbursement2008 bucket2008 diabetes ihd
## 31 20 13 13
## heart.failure kidney arthritis
## 11 9 1
##
## Node number 1: 274803 observations, complexity param=0.04908841
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (165987 obs) right son=3 (108816 obs)
## Primary splits:
## reimbursement2008 < 1565 to the left, improve=24395.14, (0 missing)
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## ihd < 0.5 to the left, improve=16291.74, (0 missing)
## diabetes < 0.5 to the left, improve=16041.26, (0 missing)
## heart.failure < 0.5 to the left, improve=12498.16, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.861, adj=0.650, (0 split)
## ihd < 0.5 to the left, agree=0.792, adj=0.474, (0 split)
## diabetes < 0.5 to the left, agree=0.785, adj=0.456, (0 split)
## heart.failure < 0.5 to the left, agree=0.762, adj=0.399, (0 split)
## kidney < 0.5 to the left, agree=0.731, adj=0.321, (0 split)
##
## Node number 2: 165987 observations
## predicted class=B1 expected loss=0.1261424 P(node) =0.6040218
## class counts: 145049 12284 6102 2315 237
## probabilities: 0.874 0.074 0.037 0.014 0.001
##
## Node number 3: 108816 observations, complexity param=0.04908841
## predicted class=B2 expected loss=0.6326367 P(node) =0.3959782
## class counts: 39417 39975 18484 9591 1349
## probabilities: 0.362 0.367 0.170 0.088 0.012
## left son=6 (39298 obs) right son=7 (69518 obs)
## Primary splits:
## reimbursement2008 < 3065 to the left, improve=2010.3080, (0 missing)
## bucket2008 < 1.5 to the left, improve=1980.9770, (0 missing)
## kidney < 0.5 to the left, improve=1416.9220, (0 missing)
## diabetes < 0.5 to the left, improve=1236.1460, (0 missing)
## heart.failure < 0.5 to the left, improve= 976.9427, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.989, adj=0.969, (0 split)
## ihd < 0.5 to the left, agree=0.659, adj=0.056, (0 split)
## diabetes < 0.5 to the left, agree=0.641, adj=0.006, (0 split)
##
## Node number 6: 39298 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.4797445 P(node) =0.1430043
## class counts: 20445 12134 4756 1782 181
## probabilities: 0.520 0.309 0.121 0.045 0.005
## left son=12 (20077 obs) right son=13 (19221 obs)
## Primary splits:
## reimbursement2008 < 2175 to the left, improve=192.7592, (0 missing)
## diabetes < 0.5 to the left, improve=155.3521, (0 missing)
## ihd < 0.5 to the left, improve=114.8541, (0 missing)
## arthritis < 0.5 to the left, improve=114.6837, (0 missing)
## kidney < 0.5 to the left, improve=108.9096, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.542, adj=0.063, (0 split)
## arthritis < 0.5 to the left, agree=0.539, adj=0.058, (0 split)
## ihd < 0.5 to the left, agree=0.534, adj=0.048, (0 split)
## kidney < 0.5 to the left, agree=0.532, adj=0.044, (0 split)
## diabetes < 0.5 to the left, agree=0.532, adj=0.043, (0 split)
##
## Node number 7: 69518 observations, complexity param=0.01395884
## predicted class=B2 expected loss=0.5995138 P(node) =0.2529739
## class counts: 18972 27841 13728 7809 1168
## probabilities: 0.273 0.400 0.197 0.112 0.017
## left son=14 (15717 obs) right son=15 (53801 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=646.4740, (0 missing)
## kidney < 0.5 to the left, improve=604.0313, (0 missing)
## arthritis < 0.5 to the left, improve=501.1263, (0 missing)
## ihd < 0.5 to the left, improve=427.9009, (0 missing)
## heart.failure < 0.5 to the left, improve=380.0080, (0 missing)
##
## Node number 12: 20077 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.4247148 P(node) =0.07305961
## class counts: 11550 5416 2200 834 77
## probabilities: 0.575 0.270 0.110 0.042 0.004
## left son=24 (8826 obs) right son=25 (11251 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=62.34344, (0 missing)
## kidney < 0.5 to the left, improve=42.15624, (0 missing)
## ihd < 0.5 to the left, improve=40.01287, (0 missing)
## heart.failure < 0.5 to the left, improve=36.00697, (0 missing)
## arthritis < 0.5 to the left, improve=33.77686, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.588, adj=0.062, (0 split)
##
## Node number 13: 19221 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.5372249 P(node) =0.06994465
## class counts: 8895 6718 2556 948 104
## probabilities: 0.463 0.350 0.133 0.049 0.005
## left son=26 (7137 obs) right son=27 (12084 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=71.31724, (0 missing)
## arthritis < 0.5 to the left, improve=61.00585, (0 missing)
## ihd < 0.5 to the left, improve=55.20411, (0 missing)
## heart.failure < 0.5 to the left, improve=52.20163, (0 missing)
## kidney < 0.5 to the left, improve=49.73230, (0 missing)
##
## Node number 14: 15717 observations, complexity param=0.004350377
## predicted class=B1 expected loss=0.5704651 P(node) =0.0571937
## class counts: 6751 5490 2365 999 112
## probabilities: 0.430 0.349 0.150 0.064 0.007
## left son=28 (13123 obs) right son=29 (2594 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=130.12270, (0 missing)
## arthritis < 0.5 to the left, improve=125.41530, (0 missing)
## ihd < 0.5 to the left, improve= 80.76118, (0 missing)
## depression < 0.5 to the left, improve= 61.32779, (0 missing)
## osteoporosis < 0.5 to the left, improve= 44.50253, (0 missing)
##
## Node number 15: 53801 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.5845616 P(node) =0.1957802
## class counts: 12221 22351 11363 6810 1056
## probabilities: 0.227 0.415 0.211 0.127 0.020
## left son=30 (25067 obs) right son=31 (28734 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=408.9756, (0 missing)
## reimbursement2008 < 15395 to the left, improve=327.1281, (0 missing)
## bucket2008 < 3.5 to the left, improve=313.8191, (0 missing)
## arthritis < 0.5 to the left, improve=266.5595, (0 missing)
## heart.failure < 0.5 to the left, improve=209.4718, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8365 to the left, agree=0.666, adj=0.282, (0 split)
## bucket2008 < 2.5 to the left, agree=0.664, adj=0.279, (0 split)
## heart.failure < 0.5 to the left, agree=0.628, adj=0.201, (0 split)
## copd < 0.5 to the left, agree=0.595, adj=0.132, (0 split)
## ihd < 0.5 to the left, agree=0.575, adj=0.089, (0 split)
##
## Node number 24: 8826 observations
## predicted class=B1 expected loss=0.3716293 P(node) =0.03211755
## class counts: 5546 2137 805 312 26
## probabilities: 0.628 0.242 0.091 0.035 0.003
##
## Node number 25: 11251 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.4663585 P(node) =0.04094206
## class counts: 6004 3279 1395 522 51
## probabilities: 0.534 0.291 0.124 0.046 0.005
## left son=50 (9007 obs) right son=51 (2244 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=18.86048, (0 missing)
## heart.failure < 0.5 to the left, improve=17.29926, (0 missing)
## arthritis < 0.5 to the left, improve=16.91283, (0 missing)
## reimbursement2008 < 1875 to the left, improve=16.48954, (0 missing)
## cancer < 0.5 to the left, improve=14.98495, (0 missing)
##
## Node number 26: 7137 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.470786 P(node) =0.02597133
## class counts: 3777 2233 794 300 33
## probabilities: 0.529 0.313 0.111 0.042 0.005
## left son=52 (5554 obs) right son=53 (1583 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=24.840370, (0 missing)
## depression < 0.5 to the left, improve=16.217060, (0 missing)
## ihd < 0.5 to the left, improve=13.895180, (0 missing)
## copd < 0.5 to the left, improve=12.688930, (0 missing)
## kidney < 0.5 to the left, improve= 9.728645, (0 missing)
##
## Node number 27: 12084 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.5764647 P(node) =0.04397332
## class counts: 5118 4485 1762 648 71
## probabilities: 0.424 0.371 0.146 0.054 0.006
## left son=54 (8413 obs) right son=55 (3671 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=27.83165, (0 missing)
## heart.failure < 0.5 to the left, improve=26.70933, (0 missing)
## ihd < 0.5 to the left, improve=24.37311, (0 missing)
## kidney < 0.5 to the left, improve=22.60183, (0 missing)
## reimbursement2008 < 2655 to the left, improve=21.75660, (0 missing)
##
## Node number 28: 13123 observations, complexity param=0.00359764
## predicted class=B1 expected loss=0.5360055 P(node) =0.04775421
## class counts: 6089 4435 1751 763 85
## probabilities: 0.464 0.338 0.133 0.058 0.006
## left son=56 (9625 obs) right son=57 (3498 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=126.28480, (0 missing)
## ihd < 0.5 to the left, improve= 70.76778, (0 missing)
## depression < 0.5 to the left, improve= 68.94332, (0 missing)
## osteoporosis < 0.5 to the left, improve= 46.31934, (0 missing)
## heart.failure < 0.5 to the left, improve= 30.26771, (0 missing)
##
## Node number 29: 2594 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.5932922 P(node) =0.009439489
## class counts: 662 1055 614 236 27
## probabilities: 0.255 0.407 0.237 0.091 0.010
## left son=58 (1000 obs) right son=59 (1594 obs)
## Primary splits:
## reimbursement2008 < 5770 to the left, improve=8.464458, (0 missing)
## arthritis < 0.5 to the left, improve=7.371565, (0 missing)
## ihd < 0.5 to the left, improve=5.410820, (0 missing)
## copd < 0.5 to the left, improve=5.301788, (0 missing)
## heart.failure < 0.5 to the left, improve=3.070575, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.823, adj=0.542, (0 split)
##
## Node number 30: 25067 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.57091 P(node) =0.09121807
## class counts: 7517 10756 4691 1917 186
## probabilities: 0.300 0.429 0.187 0.076 0.007
## left son=60 (15178 obs) right son=61 (9889 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=169.25970, (0 missing)
## cancer < 0.5 to the left, improve= 99.57556, (0 missing)
## ihd < 0.5 to the left, improve= 68.28883, (0 missing)
## depression < 0.5 to the left, improve= 61.94482, (0 missing)
## heart.failure < 0.5 to the left, improve= 42.19646, (0 missing)
## Surrogate splits:
## reimbursement2008 < 66495 to the left, agree=0.606, adj=0.001, (0 split)
##
## Node number 31: 28734 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.5964711 P(node) =0.1045622
## class counts: 4704 11595 6672 4893 870
## probabilities: 0.164 0.404 0.232 0.170 0.030
## left son=62 (16249 obs) right son=63 (12485 obs)
## Primary splits:
## reimbursement2008 < 15395 to the left, improve=177.49270, (0 missing)
## bucket2008 < 3.5 to the left, improve=170.28940, (0 missing)
## arthritis < 0.5 to the left, improve=101.31920, (0 missing)
## heart.failure < 0.5 to the left, improve= 62.82321, (0 missing)
## ihd < 0.5 to the left, improve= 55.35075, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.924, adj=0.826, (0 split)
## copd < 0.5 to the left, agree=0.609, adj=0.101, (0 split)
## stroke < 0.5 to the left, agree=0.605, adj=0.091, (0 split)
## cancer < 0.5 to the left, agree=0.580, adj=0.033, (0 split)
## alzheimers < 0.5 to the left, agree=0.569, adj=0.008, (0 split)
##
## Node number 50: 9007 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.4490951 P(node) =0.03277621
## class counts: 4962 2540 1087 378 40
## probabilities: 0.551 0.282 0.121 0.042 0.004
## left son=100 (4935 obs) right son=101 (4072 obs)
## Primary splits:
## reimbursement2008 < 1875 to the left, improve=14.670650, (0 missing)
## cancer < 0.5 to the left, improve=12.077140, (0 missing)
## arthritis < 0.5 to the left, improve= 9.470091, (0 missing)
## heart.failure < 0.5 to the left, improve= 7.308909, (0 missing)
## depression < 0.5 to the left, improve= 6.801973, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.549, adj=0.003, (0 split)
## age < 29.5 to the right, agree=0.548, adj=0.001, (0 split)
##
## Node number 51: 2244 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5356506 P(node) =0.00816585
## class counts: 1042 739 308 144 11
## probabilities: 0.464 0.329 0.137 0.064 0.005
## left son=102 (992 obs) right son=103 (1252 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=7.795458, (0 missing)
## arthritis < 0.5 to the left, improve=7.027320, (0 missing)
## ihd < 0.5 to the left, improve=4.964222, (0 missing)
## reimbursement2008 < 1735 to the left, improve=4.132280, (0 missing)
## cancer < 0.5 to the left, improve=3.835396, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.565, adj=0.016, (0 split)
## age < 33.5 to the left, agree=0.559, adj=0.002, (0 split)
##
## Node number 52: 5554 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4449046 P(node) =0.02021084
## class counts: 3083 1647 580 217 27
## probabilities: 0.555 0.297 0.104 0.039 0.005
## left son=104 (2348 obs) right son=105 (3206 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=13.118310, (0 missing)
## depression < 0.5 to the left, improve=12.689550, (0 missing)
## kidney < 0.5 to the left, improve= 9.684755, (0 missing)
## copd < 0.5 to the left, improve= 9.145592, (0 missing)
## heart.failure < 0.5 to the left, improve= 8.228139, (0 missing)
## Surrogate splits:
## age < 28.5 to the left, agree=0.579, adj=0.004, (0 split)
## reimbursement2008 < 2185 to the left, agree=0.578, adj=0.001, (0 split)
##
## Node number 53: 1583 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5615919 P(node) =0.00576049
## class counts: 694 586 214 83 6
## probabilities: 0.438 0.370 0.135 0.052 0.004
## left son=106 (1525 obs) right son=107 (58 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=5.133391, (0 missing)
## reimbursement2008 < 2725 to the left, improve=3.164238, (0 missing)
## cancer < 0.5 to the left, improve=2.451745, (0 missing)
## copd < 0.5 to the left, improve=2.436381, (0 missing)
## depression < 0.5 to the left, improve=1.979459, (0 missing)
##
## Node number 54: 8413 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.5530726 P(node) =0.03061466
## class counts: 3760 2943 1225 438 47
## probabilities: 0.447 0.350 0.146 0.052 0.006
## left son=108 (4375 obs) right son=109 (4038 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=25.12070, (0 missing)
## ihd < 0.5 to the left, improve=19.50225, (0 missing)
## kidney < 0.5 to the left, improve=18.23799, (0 missing)
## depression < 0.5 to the left, improve=14.07225, (0 missing)
## reimbursement2008 < 2615 to the left, improve=12.21338, (0 missing)
## Surrogate splits:
## kidney < 0.5 to the left, agree=0.569, adj=0.103, (0 split)
## copd < 0.5 to the left, agree=0.568, adj=0.100, (0 split)
## alzheimers < 0.5 to the left, agree=0.546, adj=0.054, (0 split)
## ihd < 0.5 to the left, agree=0.544, adj=0.050, (0 split)
## stroke < 0.5 to the left, agree=0.536, adj=0.034, (0 split)
##
## Node number 55: 3671 observations, complexity param=0.0002988809
## predicted class=B2 expected loss=0.579951 P(node) =0.01335866
## class counts: 1358 1542 537 210 24
## probabilities: 0.370 0.420 0.146 0.057 0.007
## left son=110 (2068 obs) right son=111 (1603 obs)
## Primary splits:
## reimbursement2008 < 2665 to the left, improve=10.442080, (0 missing)
## cancer < 0.5 to the left, improve= 4.234333, (0 missing)
## ihd < 0.5 to the left, improve= 4.129116, (0 missing)
## kidney < 0.5 to the left, improve= 3.679214, (0 missing)
## copd < 0.5 to the left, improve= 3.281268, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.638, adj=0.170, (0 split)
## cancer < 0.5 to the left, agree=0.566, adj=0.006, (0 split)
## age < 26.5 to the right, agree=0.564, adj=0.001, (0 split)
## stroke < 0.5 to the left, agree=0.564, adj=0.001, (0 split)
##
## Node number 56: 9625 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.4874805 P(node) =0.03502509
## class counts: 4933 2954 1162 520 56
## probabilities: 0.513 0.307 0.121 0.054 0.006
## left son=112 (3135 obs) right son=113 (6490 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=51.18602, (0 missing)
## depression < 0.5 to the left, improve=46.82343, (0 missing)
## heart.failure < 0.5 to the left, improve=27.25528, (0 missing)
## osteoporosis < 0.5 to the left, improve=25.54800, (0 missing)
## reimbursement2008 < 6615 to the left, improve=12.84564, (0 missing)
##
## Node number 57: 3498 observations, complexity param=0.0004372516
## predicted class=B2 expected loss=0.5766152 P(node) =0.01272912
## class counts: 1156 1481 589 243 29
## probabilities: 0.330 0.423 0.168 0.069 0.008
## left son=114 (2340 obs) right son=115 (1158 obs)
## Primary splits:
## reimbursement2008 < 8525 to the left, improve=12.263650, (0 missing)
## depression < 0.5 to the left, improve=10.454350, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.052395, (0 missing)
## copd < 0.5 to the left, improve= 8.848663, (0 missing)
## ihd < 0.5 to the left, improve= 8.087092, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.979, adj=0.935, (0 split)
## kidney < 0.5 to the left, agree=0.692, adj=0.069, (0 split)
## stroke < 0.5 to the left, agree=0.680, adj=0.033, (0 split)
##
## Node number 58: 1000 observations
## predicted class=B2 expected loss=0.562 P(node) =0.00363897
## class counts: 296 438 191 70 5
## probabilities: 0.296 0.438 0.191 0.070 0.005
##
## Node number 59: 1594 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6129235 P(node) =0.005800519
## class counts: 366 617 423 166 22
## probabilities: 0.230 0.387 0.265 0.104 0.014
## left son=118 (1054 obs) right son=119 (540 obs)
## Primary splits:
## reimbursement2008 < 8645 to the right, improve=7.014383, (0 missing)
## arthritis < 0.5 to the left, improve=5.636989, (0 missing)
## bucket2008 < 2.5 to the right, improve=4.256675, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.245615, (0 missing)
## ihd < 0.5 to the left, improve=2.672736, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.949, adj=0.848, (0 split)
## age < 27.5 to the right, agree=0.662, adj=0.002, (0 split)
##
## Node number 60: 15178 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.6047569 P(node) =0.05523229
## class counts: 5388 5999 2649 1047 95
## probabilities: 0.355 0.395 0.175 0.069 0.006
## left son=120 (12572 obs) right son=121 (2606 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=92.65854, (0 missing)
## ihd < 0.5 to the left, improve=43.72992, (0 missing)
## depression < 0.5 to the left, improve=36.05906, (0 missing)
## heart.failure < 0.5 to the left, improve=30.26654, (0 missing)
## copd < 0.5 to the left, improve=25.73984, (0 missing)
##
## Node number 61: 9889 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5189605 P(node) =0.03598578
## class counts: 2129 4757 2042 870 91
## probabilities: 0.215 0.481 0.206 0.088 0.009
## left son=122 (5134 obs) right son=123 (4755 obs)
## Primary splits:
## depression < 0.5 to the left, improve=18.84327, (0 missing)
## cancer < 0.5 to the left, improve=17.45891, (0 missing)
## ihd < 0.5 to the left, improve=13.35120, (0 missing)
## reimbursement2008 < 9795 to the left, improve=12.33086, (0 missing)
## copd < 0.5 to the left, improve=12.26415, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.564, adj=0.093, (0 split)
## copd < 0.5 to the left, agree=0.546, adj=0.056, (0 split)
## reimbursement2008 < 5815 to the left, agree=0.542, adj=0.048, (0 split)
## age < 64.5 to the right, agree=0.537, adj=0.037, (0 split)
## bucket2008 < 2.5 to the left, agree=0.536, adj=0.036, (0 split)
##
## Node number 62: 16249 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.5619423 P(node) =0.05912963
## class counts: 3113 7118 3819 1946 253
## probabilities: 0.192 0.438 0.235 0.120 0.016
## left son=124 (9424 obs) right son=125 (6825 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=70.60653, (0 missing)
## cancer < 0.5 to the left, improve=30.24922, (0 missing)
## ihd < 0.5 to the left, improve=29.86941, (0 missing)
## reimbursement2008 < 5665 to the left, improve=23.89268, (0 missing)
## bucket2008 < 2.5 to the left, improve=21.55872, (0 missing)
##
## Node number 63: 12485 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.6414097 P(node) =0.04543255
## class counts: 1591 4477 2853 2947 617
## probabilities: 0.127 0.359 0.229 0.236 0.049
## left son=126 (5402 obs) right son=127 (7083 obs)
## Primary splits:
## arthritis < 0.5 to the right, improve=35.40534, (0 missing)
## cancer < 0.5 to the left, improve=26.78171, (0 missing)
## reimbursement2008 < 26625 to the left, improve=24.60405, (0 missing)
## depression < 0.5 to the left, improve=23.29796, (0 missing)
## heart.failure < 0.5 to the left, improve=17.01274, (0 missing)
## Surrogate splits:
## age < 28.5 to the left, agree=0.568, adj=0.002, (0 split)
## reimbursement2008 < 15435 to the left, agree=0.568, adj=0.001, (0 split)
##
## Node number 100: 4935 observations
## predicted class=B1 expected loss=0.4196555 P(node) =0.01795832
## class counts: 2864 1294 550 205 22
## probabilities: 0.580 0.262 0.111 0.042 0.004
##
## Node number 101: 4072 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.4847741 P(node) =0.01481789
## class counts: 2098 1246 537 173 18
## probabilities: 0.515 0.306 0.132 0.042 0.004
## left son=202 (3786 obs) right son=203 (286 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=5.937439, (0 missing)
## arthritis < 0.5 to the left, improve=5.625805, (0 missing)
## copd < 0.5 to the left, improve=3.348444, (0 missing)
## ihd < 0.5 to the left, improve=3.030239, (0 missing)
## heart.failure < 0.5 to the left, improve=2.851779, (0 missing)
##
## Node number 102: 992 observations
## predicted class=B1 expected loss=0.4808468 P(node) =0.003609859
## class counts: 515 292 126 57 2
## probabilities: 0.519 0.294 0.127 0.057 0.002
##
## Node number 103: 1252 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5790735 P(node) =0.004555991
## class counts: 527 447 182 87 9
## probabilities: 0.421 0.357 0.145 0.069 0.007
## left son=206 (904 obs) right son=207 (348 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=7.739842, (0 missing)
## age < 93.5 to the left, improve=3.754099, (0 missing)
## cancer < 0.5 to the left, improve=3.514161, (0 missing)
## reimbursement2008 < 1955 to the left, improve=3.377454, (0 missing)
## ihd < 0.5 to the left, improve=1.751139, (0 missing)
## Surrogate splits:
## age < 30.5 to the right, agree=0.724, adj=0.006, (0 split)
##
## Node number 104: 2348 observations
## predicted class=B1 expected loss=0.3973595 P(node) =0.008544303
## class counts: 1415 632 217 72 12
## probabilities: 0.603 0.269 0.092 0.031 0.005
##
## Node number 105: 3206 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4797255 P(node) =0.01166654
## class counts: 1668 1015 363 145 15
## probabilities: 0.520 0.317 0.113 0.045 0.005
## left son=210 (2325 obs) right son=211 (881 obs)
## Primary splits:
## depression < 0.5 to the left, improve=8.135493, (0 missing)
## kidney < 0.5 to the left, improve=5.219511, (0 missing)
## reimbursement2008 < 2785 to the left, improve=4.205524, (0 missing)
## heart.failure < 0.5 to the left, improve=3.201394, (0 missing)
## copd < 0.5 to the left, improve=3.002159, (0 missing)
## Surrogate splits:
## age < 29.5 to the right, agree=0.726, adj=0.003, (0 split)
##
## Node number 106: 1525 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5534426 P(node) =0.00554943
## class counts: 681 554 202 82 6
## probabilities: 0.447 0.363 0.132 0.054 0.004
## left son=212 (1438 obs) right son=213 (87 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.548424, (0 missing)
## reimbursement2008 < 2715 to the right, improve=2.513748, (0 missing)
## copd < 0.5 to the left, improve=1.973703, (0 missing)
## depression < 0.5 to the left, improve=1.853940, (0 missing)
## kidney < 0.5 to the left, improve=1.632947, (0 missing)
##
## Node number 107: 58 observations
## predicted class=B2 expected loss=0.4482759 P(node) =0.0002110603
## class counts: 13 32 12 1 0
## probabilities: 0.224 0.552 0.207 0.017 0.000
##
## Node number 108: 4375 observations, complexity param=0.0002435326
## predicted class=B1 expected loss=0.5074286 P(node) =0.0159205
## class counts: 2155 1478 555 170 17
## probabilities: 0.493 0.338 0.127 0.039 0.004
## left son=216 (3992 obs) right son=217 (383 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=10.015540, (0 missing)
## ihd < 0.5 to the left, improve= 9.488719, (0 missing)
## depression < 0.5 to the left, improve= 7.316301, (0 missing)
## reimbursement2008 < 2615 to the left, improve= 5.949976, (0 missing)
## copd < 0.5 to the left, improve= 5.117423, (0 missing)
##
## Node number 109: 4038 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.602526 P(node) =0.01469416
## class counts: 1605 1465 670 268 30
## probabilities: 0.397 0.363 0.166 0.066 0.007
## left son=218 (2819 obs) right son=219 (1219 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=10.392200, (0 missing)
## reimbursement2008 < 2455 to the left, improve= 6.028802, (0 missing)
## ihd < 0.5 to the left, improve= 5.795095, (0 missing)
## depression < 0.5 to the left, improve= 5.214940, (0 missing)
## stroke < 0.5 to the left, improve= 3.343262, (0 missing)
##
## Node number 110: 2068 observations, complexity param=0.0002767415
## predicted class=B1 expected loss=0.5918762 P(node) =0.007525391
## class counts: 844 817 280 117 10
## probabilities: 0.408 0.395 0.135 0.057 0.005
## left son=220 (517 obs) right son=221 (1551 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.581883, (0 missing)
## reimbursement2008 < 2305 to the left, improve=3.255344, (0 missing)
## cancer < 0.5 to the left, improve=3.097089, (0 missing)
## age < 54.5 to the left, improve=1.964830, (0 missing)
## kidney < 0.5 to the left, improve=1.730688, (0 missing)
##
## Node number 111: 1603 observations
## predicted class=B2 expected loss=0.547723 P(node) =0.00583327
## class counts: 514 725 257 93 14
## probabilities: 0.321 0.452 0.160 0.058 0.009
##
## Node number 112: 3135 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.3974482 P(node) =0.01140817
## class counts: 1889 825 298 113 10
## probabilities: 0.603 0.263 0.095 0.036 0.003
## left son=224 (2292 obs) right son=225 (843 obs)
## Primary splits:
## depression < 0.5 to the left, improve=19.892930, (0 missing)
## reimbursement2008 < 9505 to the right, improve=15.211730, (0 missing)
## bucket2008 < 2.5 to the right, improve=13.054300, (0 missing)
## osteoporosis < 0.5 to the left, improve=10.317040, (0 missing)
## age < 92.5 to the left, improve= 3.244996, (0 missing)
## Surrogate splits:
## reimbursement2008 < 60755 to the left, agree=0.731, adj=0.001, (0 split)
##
## Node number 113: 6490 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.5309707 P(node) =0.02361692
## class counts: 3044 2129 864 407 46
## probabilities: 0.469 0.328 0.133 0.063 0.007
## left son=226 (4266 obs) right son=227 (2224 obs)
## Primary splits:
## depression < 0.5 to the left, improve=22.130520, (0 missing)
## heart.failure < 0.5 to the left, improve=12.472230, (0 missing)
## osteoporosis < 0.5 to the left, improve=12.135520, (0 missing)
## reimbursement2008 < 6615 to the left, improve=10.028930, (0 missing)
## bucket2008 < 2.5 to the left, improve= 8.000565, (0 missing)
## Surrogate splits:
## age < 34.5 to the right, agree=0.658, adj=0.003, (0 split)
## reimbursement2008 < 115145 to the left, agree=0.658, adj=0.001, (0 split)
##
## Node number 114: 2340 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.542735 P(node) =0.008515191
## class counts: 720 1070 391 144 15
## probabilities: 0.308 0.457 0.167 0.062 0.006
## left son=228 (1359 obs) right son=229 (981 obs)
## Primary splits:
## reimbursement2008 < 4645 to the left, improve=5.782135, (0 missing)
## ihd < 0.5 to the left, improve=5.431632, (0 missing)
## depression < 0.5 to the left, improve=4.505952, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.336155, (0 missing)
## alzheimers < 0.5 to the left, improve=3.247654, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.613, adj=0.076, (0 split)
## copd < 0.5 to the left, agree=0.606, adj=0.059, (0 split)
## kidney < 0.5 to the left, agree=0.586, adj=0.013, (0 split)
## age < 91.5 to the left, agree=0.585, adj=0.011, (0 split)
## stroke < 0.5 to the left, agree=0.585, adj=0.009, (0 split)
##
## Node number 115: 1158 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.6234888 P(node) =0.004213928
## class counts: 436 411 198 99 14
## probabilities: 0.377 0.355 0.171 0.085 0.012
## left son=230 (714 obs) right son=231 (444 obs)
## Primary splits:
## copd < 0.5 to the left, improve=13.168040, (0 missing)
## depression < 0.5 to the left, improve= 8.948306, (0 missing)
## kidney < 0.5 to the left, improve= 6.276303, (0 missing)
## ihd < 0.5 to the left, improve= 5.293866, (0 missing)
## reimbursement2008 < 14980 to the left, improve= 4.056180, (0 missing)
## Surrogate splits:
## age < 94.5 to the left, agree=0.626, adj=0.025, (0 split)
## reimbursement2008 < 72745 to the left, agree=0.620, adj=0.009, (0 split)
##
## Node number 118: 1054 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6223909 P(node) =0.003835475
## class counts: 281 398 250 109 16
## probabilities: 0.267 0.378 0.237 0.103 0.015
## left son=236 (745 obs) right son=237 (309 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=8.492364, (0 missing)
## ihd < 0.5 to the left, improve=3.739184, (0 missing)
## depression < 0.5 to the left, improve=2.714506, (0 missing)
## copd < 0.5 to the left, improve=2.704564, (0 missing)
## reimbursement2008 < 67610 to the left, improve=2.665770, (0 missing)
## Surrogate splits:
## age < 29.5 to the right, agree=0.708, adj=0.003, (0 split)
##
## Node number 119: 540 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.5944444 P(node) =0.001965044
## class counts: 85 219 173 57 6
## probabilities: 0.157 0.406 0.320 0.106 0.011
## left son=238 (243 obs) right son=239 (297 obs)
## Primary splits:
## heart.failure < 0.5 to the right, improve=3.144781, (0 missing)
## reimbursement2008 < 7455 to the left, improve=1.665302, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.352183, (0 missing)
## age < 86.5 to the right, improve=1.232072, (0 missing)
## arthritis < 0.5 to the right, improve=1.028824, (0 missing)
## Surrogate splits:
## copd < 0.5 to the right, agree=0.604, adj=0.119, (0 split)
## kidney < 0.5 to the right, agree=0.585, adj=0.078, (0 split)
## stroke < 0.5 to the right, agree=0.583, adj=0.074, (0 split)
## arthritis < 0.5 to the right, agree=0.576, adj=0.058, (0 split)
## depression < 0.5 to the right, agree=0.572, adj=0.049, (0 split)
##
## Node number 120: 12572 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.613188 P(node) =0.04574914
## class counts: 4844 4863 2000 791 74
## probabilities: 0.385 0.387 0.159 0.063 0.006
## left son=240 (2617 obs) right son=241 (9955 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=36.80981, (0 missing)
## depression < 0.5 to the left, improve=36.47326, (0 missing)
## heart.failure < 0.5 to the left, improve=27.52215, (0 missing)
## copd < 0.5 to the left, improve=21.85222, (0 missing)
## reimbursement2008 < 8955 to the left, improve=19.34797, (0 missing)
##
## Node number 121: 2606 observations
## predicted class=B2 expected loss=0.5640829 P(node) =0.009483157
## class counts: 544 1136 649 256 21
## probabilities: 0.209 0.436 0.249 0.098 0.008
##
## Node number 122: 5134 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5190884 P(node) =0.01868247
## class counts: 1277 2469 936 412 40
## probabilities: 0.249 0.481 0.182 0.080 0.008
## left son=244 (4305 obs) right son=245 (829 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=12.348810, (0 missing)
## reimbursement2008 < 9985 to the left, improve=11.590150, (0 missing)
## bucket2008 < 2.5 to the left, improve= 7.979608, (0 missing)
## ihd < 0.5 to the left, improve= 7.512372, (0 missing)
## copd < 0.5 to the left, improve= 7.186891, (0 missing)
##
## Node number 123: 4755 observations
## predicted class=B2 expected loss=0.5188223 P(node) =0.0173033
## class counts: 852 2288 1106 458 51
## probabilities: 0.179 0.481 0.233 0.096 0.011
##
## Node number 124: 9424 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.5992148 P(node) =0.03429366
## class counts: 2192 3777 2139 1156 160
## probabilities: 0.233 0.401 0.227 0.123 0.017
## left son=248 (7786 obs) right son=249 (1638 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=31.06099, (0 missing)
## ihd < 0.5 to the left, improve=19.93184, (0 missing)
## depression < 0.5 to the left, improve=16.57581, (0 missing)
## reimbursement2008 < 6325 to the left, improve=12.91187, (0 missing)
## bucket2008 < 2.5 to the left, improve=10.82187, (0 missing)
##
## Node number 125: 6825 observations
## predicted class=B2 expected loss=0.5104762 P(node) =0.02483597
## class counts: 921 3341 1680 790 93
## probabilities: 0.135 0.490 0.246 0.116 0.014
##
## Node number 126: 5402 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5960755 P(node) =0.01965772
## class counts: 509 2182 1310 1186 215
## probabilities: 0.094 0.404 0.243 0.220 0.040
## left son=252 (3345 obs) right son=253 (2057 obs)
## Primary splits:
## reimbursement2008 < 34925 to the left, improve=14.212070, (0 missing)
## copd < 0.5 to the left, improve=10.384850, (0 missing)
## depression < 0.5 to the left, improve= 8.104595, (0 missing)
## cancer < 0.5 to the right, improve= 6.743072, (0 missing)
## heart.failure < 0.5 to the left, improve= 6.417519, (0 missing)
## Surrogate splits:
## bucket2008 < 4.5 to the left, agree=0.776, adj=0.413, (0 split)
##
## Node number 127: 7083 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.6759848 P(node) =0.02577483
## class counts: 1082 2295 1543 1761 402
## probabilities: 0.153 0.324 0.218 0.249 0.057
## left son=254 (5298 obs) right son=255 (1785 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=21.09129, (0 missing)
## depression < 0.5 to the left, improve=19.29947, (0 missing)
## reimbursement2008 < 26625 to the left, improve=15.18952, (0 missing)
## copd < 0.5 to the left, improve=14.68870, (0 missing)
## heart.failure < 0.5 to the left, improve=12.81802, (0 missing)
##
## Node number 202: 3786 observations
## predicted class=B1 expected loss=0.4772847 P(node) =0.01377714
## class counts: 1979 1131 501 157 18
## probabilities: 0.523 0.299 0.132 0.041 0.005
##
## Node number 203: 286 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5839161 P(node) =0.001040746
## class counts: 119 115 36 16 0
## probabilities: 0.416 0.402 0.126 0.056 0.000
## left son=406 (128 obs) right son=407 (158 obs)
## Primary splits:
## age < 73.5 to the left, improve=2.9724540, (0 missing)
## reimbursement2008 < 2005 to the left, improve=1.9802050, (0 missing)
## depression < 0.5 to the left, improve=0.5460014, (0 missing)
## alzheimers < 0.5 to the right, improve=0.4144954, (0 missing)
## ihd < 0.5 to the left, improve=0.3767582, (0 missing)
## Surrogate splits:
## reimbursement2008 < 1945 to the left, agree=0.580, adj=0.063, (0 split)
## arthritis < 0.5 to the right, agree=0.563, adj=0.023, (0 split)
##
## Node number 206: 904 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5376106 P(node) =0.003289629
## class counts: 418 304 119 57 6
## probabilities: 0.462 0.336 0.132 0.063 0.007
## left son=412 (270 obs) right son=413 (634 obs)
## Primary splits:
## reimbursement2008 < 1735 to the left, improve=3.8438620, (0 missing)
## age < 93.5 to the left, improve=3.6681650, (0 missing)
## ihd < 0.5 to the left, improve=3.2669730, (0 missing)
## cancer < 0.5 to the left, improve=3.0869480, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7446912, (0 missing)
## Surrogate splits:
## age < 29 to the left, agree=0.702, adj=0.004, (0 split)
##
## Node number 207: 348 observations
## predicted class=B2 expected loss=0.5890805 P(node) =0.001266362
## class counts: 109 143 63 30 3
## probabilities: 0.313 0.411 0.181 0.086 0.009
##
## Node number 210: 2325 observations
## predicted class=B1 expected loss=0.4541935 P(node) =0.008460606
## class counts: 1269 700 245 99 12
## probabilities: 0.546 0.301 0.105 0.043 0.005
##
## Node number 211: 881 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5471056 P(node) =0.003205933
## class counts: 399 315 118 46 3
## probabilities: 0.453 0.358 0.134 0.052 0.003
## left son=422 (763 obs) right son=423 (118 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=3.122415, (0 missing)
## age < 78.5 to the right, improve=2.656467, (0 missing)
## reimbursement2008 < 2205 to the right, improve=1.600090, (0 missing)
## stroke < 0.5 to the left, improve=1.074836, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.071176, (0 missing)
##
## Node number 212: 1438 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5452017 P(node) =0.00523284
## class counts: 654 515 187 76 6
## probabilities: 0.455 0.358 0.130 0.053 0.004
## left son=424 (495 obs) right son=425 (943 obs)
## Primary splits:
## reimbursement2008 < 2715 to the right, improve=2.835023, (0 missing)
## kidney < 0.5 to the left, improve=1.879898, (0 missing)
## copd < 0.5 to the left, improve=1.857999, (0 missing)
## age < 40.5 to the right, improve=1.802592, (0 missing)
## heart.failure < 0.5 to the left, improve=1.761837, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=0.713, adj=0.166, (0 split)
## age < 28.5 to the left, agree=0.656, adj=0.002, (0 split)
##
## Node number 213: 87 observations
## predicted class=B2 expected loss=0.5517241 P(node) =0.0003165904
## class counts: 27 39 15 6 0
## probabilities: 0.310 0.448 0.172 0.069 0.000
##
## Node number 216: 3992 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.495491 P(node) =0.01452677
## class counts: 2014 1315 497 153 13
## probabilities: 0.505 0.329 0.124 0.038 0.003
## left son=432 (1265 obs) right son=433 (2727 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=7.867939, (0 missing)
## depression < 0.5 to the left, improve=6.016589, (0 missing)
## copd < 0.5 to the left, improve=5.402587, (0 missing)
## kidney < 0.5 to the left, improve=3.916699, (0 missing)
## reimbursement2008 < 2615 to the left, improve=3.836002, (0 missing)
##
## Node number 217: 383 observations, complexity param=0.0001826494
## predicted class=B2 expected loss=0.5744125 P(node) =0.001393726
## class counts: 141 163 58 17 4
## probabilities: 0.368 0.426 0.151 0.044 0.010
## left son=434 (238 obs) right son=435 (145 obs)
## Primary splits:
## reimbursement2008 < 2705 to the left, improve=4.9624930, (0 missing)
## depression < 0.5 to the left, improve=3.2303380, (0 missing)
## age < 67.5 to the right, improve=2.3511250, (0 missing)
## ihd < 0.5 to the left, improve=1.5735720, (0 missing)
## bucket2008 < 1.5 to the left, improve=0.9813303, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.681, adj=0.159, (0 split)
## age < 45.5 to the right, agree=0.624, adj=0.007, (0 split)
##
## Node number 218: 2819 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.5746719 P(node) =0.01025826
## class counts: 1199 980 439 183 18
## probabilities: 0.425 0.348 0.156 0.065 0.006
## left son=436 (635 obs) right son=437 (2184 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=6.072389, (0 missing)
## reimbursement2008 < 2325 to the left, improve=3.797765, (0 missing)
## age < 40.5 to the right, improve=3.110525, (0 missing)
## depression < 0.5 to the left, improve=2.993563, (0 missing)
## stroke < 0.5 to the left, improve=2.412511, (0 missing)
##
## Node number 219: 1219 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.6021329 P(node) =0.004435905
## class counts: 406 485 231 85 12
## probabilities: 0.333 0.398 0.189 0.070 0.010
## left son=438 (613 obs) right son=439 (606 obs)
## Primary splits:
## reimbursement2008 < 2615 to the left, improve=4.2080810, (0 missing)
## age < 98.5 to the right, improve=2.1482090, (0 missing)
## depression < 0.5 to the left, improve=1.6601240, (0 missing)
## stroke < 0.5 to the left, improve=0.8099205, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.7434054, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.579, adj=0.153, (0 split)
## depression < 0.5 to the left, agree=0.523, adj=0.041, (0 split)
## stroke < 0.5 to the left, agree=0.522, adj=0.038, (0 split)
## age < 65.5 to the right, agree=0.519, adj=0.033, (0 split)
## cancer < 0.5 to the left, agree=0.514, adj=0.021, (0 split)
##
## Node number 220: 517 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5299807 P(node) =0.001881348
## class counts: 243 191 57 25 1
## probabilities: 0.470 0.369 0.110 0.048 0.002
## left son=440 (143 obs) right son=441 (374 obs)
## Primary splits:
## reimbursement2008 < 2295 to the left, improve=6.0966680, (0 missing)
## cancer < 0.5 to the left, improve=2.5628030, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.9493160, (0 missing)
## age < 44.5 to the right, improve=1.5968610, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9005685, (0 missing)
## Surrogate splits:
## age < 98.5 to the right, agree=0.729, adj=0.021, (0 split)
##
## Node number 221: 1551 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5963894 P(node) =0.005644043
## class counts: 601 626 223 92 9
## probabilities: 0.387 0.404 0.144 0.059 0.006
## left son=442 (18 obs) right son=443 (1533 obs)
## Primary splits:
## age < 35 to the left, improve=3.0170030, (0 missing)
## kidney < 0.5 to the left, improve=2.3281310, (0 missing)
## cancer < 0.5 to the left, improve=1.5502140, (0 missing)
## stroke < 0.5 to the left, improve=1.1903410, (0 missing)
## copd < 0.5 to the left, improve=0.9727402, (0 missing)
##
## Node number 224: 2292 observations
## predicted class=B1 expected loss=0.3582024 P(node) =0.00834052
## class counts: 1471 549 183 79 10
## probabilities: 0.642 0.240 0.080 0.034 0.004
##
## Node number 225: 843 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5041518 P(node) =0.003067652
## class counts: 418 276 115 34 0
## probabilities: 0.496 0.327 0.136 0.040 0.000
## left son=450 (810 obs) right son=451 (33 obs)
## Primary splits:
## age < 92.5 to the left, improve=5.7055350, (0 missing)
## reimbursement2008 < 11540 to the right, improve=5.6370950, (0 missing)
## bucket2008 < 2.5 to the right, improve=2.9317810, (0 missing)
## stroke < 0.5 to the left, improve=0.7284423, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3506867, (0 missing)
##
## Node number 226: 4266 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.4946085 P(node) =0.01552385
## class counts: 2156 1343 503 238 26
## probabilities: 0.505 0.315 0.118 0.056 0.006
## left son=452 (3304 obs) right son=453 (962 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=10.212680, (0 missing)
## reimbursement2008 < 5905 to the right, improve= 9.673580, (0 missing)
## bucket2008 < 2.5 to the right, improve= 7.844764, (0 missing)
## heart.failure < 0.5 to the left, improve= 6.371374, (0 missing)
## age < 62.5 to the left, improve= 3.683231, (0 missing)
##
## Node number 227: 2224 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.6007194 P(node) =0.00809307
## class counts: 888 786 361 169 20
## probabilities: 0.399 0.353 0.162 0.076 0.009
## left son=454 (1518 obs) right son=455 (706 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=6.746609, (0 missing)
## heart.failure < 0.5 to the left, improve=4.569316, (0 missing)
## reimbursement2008 < 10710 to the left, improve=3.711923, (0 missing)
## age < 39.5 to the right, improve=3.285727, (0 missing)
## bucket2008 < 2.5 to the left, improve=2.661027, (0 missing)
## Surrogate splits:
## reimbursement2008 < 14380 to the left, agree=0.714, adj=0.101, (0 split)
## bucket2008 < 3.5 to the left, agree=0.708, adj=0.081, (0 split)
## age < 98.5 to the left, agree=0.684, adj=0.004, (0 split)
##
## Node number 228: 1359 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5548197 P(node) =0.004945361
## class counts: 467 605 203 76 8
## probabilities: 0.344 0.445 0.149 0.056 0.006
## left son=456 (440 obs) right son=457 (919 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.300387, (0 missing)
## alzheimers < 0.5 to the left, improve=2.001264, (0 missing)
## reimbursement2008 < 3265 to the left, improve=1.998939, (0 missing)
## depression < 0.5 to the left, improve=1.755319, (0 missing)
## heart.failure < 0.5 to the left, improve=1.574681, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3095 to the left, agree=0.678, adj=0.007, (0 split)
## age < 29.5 to the left, agree=0.678, adj=0.005, (0 split)
##
## Node number 229: 981 observations
## predicted class=B2 expected loss=0.5259939 P(node) =0.00356983
## class counts: 253 465 188 68 7
## probabilities: 0.258 0.474 0.192 0.069 0.007
##
## Node number 230: 714 observations, complexity param=0.0001881842
## predicted class=B1 expected loss=0.5546218 P(node) =0.002598225
## class counts: 318 239 91 61 5
## probabilities: 0.445 0.335 0.127 0.085 0.007
## left son=460 (412 obs) right son=461 (302 obs)
## Primary splits:
## depression < 0.5 to the left, improve=8.699660, (0 missing)
## age < 92.5 to the right, improve=3.253447, (0 missing)
## reimbursement2008 < 14980 to the left, improve=2.826720, (0 missing)
## bucket2008 < 3.5 to the left, improve=2.191697, (0 missing)
## kidney < 0.5 to the left, improve=2.037790, (0 missing)
## Surrogate splits:
## reimbursement2008 < 32685 to the left, agree=0.591, adj=0.033, (0 split)
## age < 35.5 to the right, agree=0.583, adj=0.013, (0 split)
##
## Node number 231: 444 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.6126126 P(node) =0.001615703
## class counts: 118 172 107 38 9
## probabilities: 0.266 0.387 0.241 0.086 0.020
## left son=462 (282 obs) right son=463 (162 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=3.735228, (0 missing)
## kidney < 0.5 to the left, improve=3.274615, (0 missing)
## reimbursement2008 < 68975 to the right, improve=3.185223, (0 missing)
## ihd < 0.5 to the left, improve=3.085645, (0 missing)
## age < 76.5 to the left, improve=1.652811, (0 missing)
## Surrogate splits:
## age < 95.5 to the left, agree=0.644, adj=0.025, (0 split)
## reimbursement2008 < 8635 to the right, agree=0.637, adj=0.006, (0 split)
##
## Node number 236: 745 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6228188 P(node) =0.002711033
## class counts: 232 281 150 72 10
## probabilities: 0.311 0.377 0.201 0.097 0.013
## left son=472 (159 obs) right son=473 (586 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.002920, (0 missing)
## reimbursement2008 < 58135 to the left, improve=2.259882, (0 missing)
## depression < 0.5 to the left, improve=2.111862, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.991400, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.920660, (0 missing)
##
## Node number 237: 309 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6213592 P(node) =0.001124442
## class counts: 49 117 100 37 6
## probabilities: 0.159 0.379 0.324 0.120 0.019
## left son=474 (237 obs) right son=475 (72 obs)
## Primary splits:
## reimbursement2008 < 10960 to the right, improve=2.966323, (0 missing)
## alzheimers < 0.5 to the right, improve=1.571780, (0 missing)
## age < 90.5 to the left, improve=1.407411, (0 missing)
## copd < 0.5 to the left, improve=1.306020, (0 missing)
## stroke < 0.5 to the left, improve=0.907593, (0 missing)
##
## Node number 238: 243 observations
## predicted class=B2 expected loss=0.526749 P(node) =0.0008842698
## class counts: 33 115 67 24 4
## probabilities: 0.136 0.473 0.276 0.099 0.016
##
## Node number 239: 297 observations, complexity param=6.167383e-05
## predicted class=B3 expected loss=0.6430976 P(node) =0.001080774
## class counts: 52 104 106 33 2
## probabilities: 0.175 0.350 0.357 0.111 0.007
## left son=478 (226 obs) right son=479 (71 obs)
## Primary splits:
## depression < 0.5 to the left, improve=1.480103, (0 missing)
## reimbursement2008 < 6875 to the right, improve=1.383473, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.357727, (0 missing)
## age < 54 to the left, improve=1.263809, (0 missing)
## alzheimers < 0.5 to the right, improve=1.096200, (0 missing)
##
## Node number 240: 2617 observations, complexity param=0.0001439056
## predicted class=B1 expected loss=0.5257929 P(node) =0.009523186
## class counts: 1241 884 351 127 14
## probabilities: 0.474 0.338 0.134 0.049 0.005
## left son=480 (403 obs) right son=481 (2214 obs)
## Primary splits:
## reimbursement2008 < 9400 to the right, improve=12.428110, (0 missing)
## bucket2008 < 2.5 to the right, improve= 8.843694, (0 missing)
## depression < 0.5 to the left, improve= 8.588030, (0 missing)
## osteoporosis < 0.5 to the left, improve= 8.405901, (0 missing)
## alzheimers < 0.5 to the left, improve= 4.036896, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.947, adj=0.658, (0 split)
##
## Node number 241: 9955 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.6003014 P(node) =0.03622595
## class counts: 3603 3979 1649 664 60
## probabilities: 0.362 0.400 0.166 0.067 0.006
## left son=482 (5563 obs) right son=483 (4392 obs)
## Primary splits:
## depression < 0.5 to the left, improve=24.69099, (0 missing)
## copd < 0.5 to the left, improve=17.49244, (0 missing)
## heart.failure < 0.5 to the left, improve=17.05734, (0 missing)
## reimbursement2008 < 8955 to the left, improve=14.88623, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.99202, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.574, adj=0.034, (0 split)
## age < 47.5 to the right, agree=0.565, adj=0.013, (0 split)
## copd < 0.5 to the left, agree=0.564, adj=0.013, (0 split)
## reimbursement2008 < 13565 to the left, agree=0.561, adj=0.005, (0 split)
## bucket2008 < 3.5 to the left, agree=0.559, adj=0.001, (0 split)
##
## Node number 244: 4305 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.524971 P(node) =0.01566577
## class counts: 1149 2045 746 328 37
## probabilities: 0.267 0.475 0.173 0.076 0.009
## left son=488 (1063 obs) right son=489 (3242 obs)
## Primary splits:
## reimbursement2008 < 9880 to the right, improve=11.346300, (0 missing)
## bucket2008 < 2.5 to the right, improve= 8.562449, (0 missing)
## ihd < 0.5 to the left, improve= 7.353611, (0 missing)
## copd < 0.5 to the left, improve= 6.701463, (0 missing)
## heart.failure < 0.5 to the left, improve= 3.881008, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.941, adj=0.762, (0 split)
##
## Node number 245: 829 observations
## predicted class=B2 expected loss=0.4885404 P(node) =0.003016707
## class counts: 128 424 190 84 3
## probabilities: 0.154 0.511 0.229 0.101 0.004
##
## Node number 248: 7786 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.6050604 P(node) =0.02833302
## class counts: 1982 3075 1667 929 133
## probabilities: 0.255 0.395 0.214 0.119 0.017
## left son=496 (964 obs) right son=497 (6822 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=18.914850, (0 missing)
## depression < 0.5 to the left, improve=16.457650, (0 missing)
## reimbursement2008 < 6325 to the left, improve=12.927220, (0 missing)
## osteoporosis < 0.5 to the left, improve= 9.344273, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.314433, (0 missing)
##
## Node number 249: 1638 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.005960634
## class counts: 210 702 472 227 27
## probabilities: 0.128 0.429 0.288 0.139 0.016
##
## Node number 252: 3345 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.580568 P(node) =0.01217236
## class counts: 372 1403 837 632 101
## probabilities: 0.111 0.419 0.250 0.189 0.030
## left son=504 (1291 obs) right son=505 (2054 obs)
## Primary splits:
## depression < 0.5 to the left, improve=6.733363, (0 missing)
## copd < 0.5 to the left, improve=6.399894, (0 missing)
## cancer < 0.5 to the left, improve=5.398776, (0 missing)
## heart.failure < 0.5 to the left, improve=3.401421, (0 missing)
## age < 31.5 to the right, improve=3.041832, (0 missing)
## Surrogate splits:
## reimbursement2008 < 15665 to the left, agree=0.614, adj=0.001, (0 split)
##
## Node number 253: 2057 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6212931 P(node) =0.007485362
## class counts: 137 779 473 554 114
## probabilities: 0.067 0.379 0.230 0.269 0.055
## left son=506 (520 obs) right son=507 (1537 obs)
## Primary splits:
## copd < 0.5 to the left, improve=4.741452, (0 missing)
## age < 62.5 to the right, improve=3.709690, (0 missing)
## cancer < 0.5 to the right, improve=3.631891, (0 missing)
## ihd < 0.5 to the left, improve=3.269099, (0 missing)
## heart.failure < 0.5 to the left, improve=3.168350, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the left, agree=0.751, adj=0.015, (0 split)
## age < 29 to the left, agree=0.749, adj=0.008, (0 split)
## ihd < 0.5 to the left, agree=0.749, adj=0.008, (0 split)
##
## Node number 254: 5298 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.689128 P(node) =0.01927927
## class counts: 908 1647 1051 1364 328
## probabilities: 0.171 0.311 0.198 0.257 0.062
## left son=508 (2489 obs) right son=509 (2809 obs)
## Primary splits:
## depression < 0.5 to the left, improve=18.17296, (0 missing)
## reimbursement2008 < 22335 to the left, improve=13.06444, (0 missing)
## copd < 0.5 to the left, improve=11.53148, (0 missing)
## ihd < 0.5 to the left, improve= 8.63716, (0 missing)
## heart.failure < 0.5 to the left, improve= 8.42218, (0 missing)
## Surrogate splits:
## copd < 0.5 to the left, agree=0.579, adj=0.104, (0 split)
## alzheimers < 0.5 to the left, agree=0.573, adj=0.092, (0 split)
## ihd < 0.5 to the left, agree=0.545, adj=0.033, (0 split)
## heart.failure < 0.5 to the left, agree=0.544, adj=0.030, (0 split)
## reimbursement2008 < 16955 to the left, agree=0.535, adj=0.010, (0 split)
##
## Node number 255: 1785 observations
## predicted class=B2 expected loss=0.6369748 P(node) =0.006495562
## class counts: 174 648 492 397 74
## probabilities: 0.097 0.363 0.276 0.222 0.041
##
## Node number 406: 128 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5 P(node) =0.0004657882
## class counts: 64 42 14 8 0
## probabilities: 0.500 0.328 0.109 0.062 0.000
## left son=812 (95 obs) right son=813 (33 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.3313600, (0 missing)
## reimbursement2008 < 2155 to the left, improve=2.1875000, (0 missing)
## age < 70.5 to the right, improve=1.5228130, (0 missing)
## arthritis < 0.5 to the left, improve=1.1806970, (0 missing)
## copd < 0.5 to the left, improve=0.4207762, (0 missing)
##
## Node number 407: 158 observations
## predicted class=B2 expected loss=0.5379747 P(node) =0.0005749573
## class counts: 55 73 22 8 0
## probabilities: 0.348 0.462 0.139 0.051 0.000
##
## Node number 412: 270 observations
## predicted class=B1 expected loss=0.462963 P(node) =0.000982522
## class counts: 145 73 36 15 1
## probabilities: 0.537 0.270 0.133 0.056 0.004
##
## Node number 413: 634 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5694006 P(node) =0.002307107
## class counts: 273 231 83 42 5
## probabilities: 0.431 0.364 0.131 0.066 0.008
## left son=826 (596 obs) right son=827 (38 obs)
## Primary splits:
## age < 91.5 to the left, improve=3.6059530, (0 missing)
## ihd < 0.5 to the left, improve=2.2411130, (0 missing)
## reimbursement2008 < 1765 to the right, improve=2.0115470, (0 missing)
## cancer < 0.5 to the left, improve=1.8824720, (0 missing)
## depression < 0.5 to the right, improve=0.5863526, (0 missing)
##
## Node number 422: 763 observations
## predicted class=B1 expected loss=0.5307995 P(node) =0.002776534
## class counts: 358 260 102 40 3
## probabilities: 0.469 0.341 0.134 0.052 0.004
##
## Node number 423: 118 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5338983 P(node) =0.0004293985
## class counts: 41 55 16 6 0
## probabilities: 0.347 0.466 0.136 0.051 0.000
## left son=846 (22 obs) right son=847 (96 obs)
## Primary splits:
## reimbursement2008 < 2865 to the right, improve=2.6611130, (0 missing)
## copd < 0.5 to the left, improve=1.5528850, (0 missing)
## heart.failure < 0.5 to the left, improve=1.3108310, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.2553930, (0 missing)
## age < 89.5 to the left, improve=0.9696791, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=0.873, adj=0.318, (0 split)
##
## Node number 424: 495 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5656566 P(node) =0.00180129
## class counts: 215 202 50 26 2
## probabilities: 0.434 0.408 0.101 0.053 0.004
## left son=848 (385 obs) right son=849 (110 obs)
## Primary splits:
## reimbursement2008 < 2795 to the right, improve=2.2427130, (0 missing)
## age < 73.5 to the left, improve=1.5903260, (0 missing)
## ihd < 0.5 to the left, improve=1.1717170, (0 missing)
## depression < 0.5 to the left, improve=0.5724615, (0 missing)
## kidney < 0.5 to the left, improve=0.5572971, (0 missing)
##
## Node number 425: 943 observations
## predicted class=B1 expected loss=0.5344645 P(node) =0.003431549
## class counts: 439 313 137 50 4
## probabilities: 0.466 0.332 0.145 0.053 0.004
##
## Node number 432: 1265 observations
## predicted class=B1 expected loss=0.4442688 P(node) =0.004603298
## class counts: 703 367 147 44 4
## probabilities: 0.556 0.290 0.116 0.035 0.003
##
## Node number 433: 2727 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5192519 P(node) =0.009923472
## class counts: 1311 948 350 109 9
## probabilities: 0.481 0.348 0.128 0.040 0.003
## left son=866 (1499 obs) right son=867 (1228 obs)
## Primary splits:
## reimbursement2008 < 2615 to the left, improve=4.028460, (0 missing)
## age < 54.5 to the left, improve=3.426946, (0 missing)
## copd < 0.5 to the left, improve=2.215284, (0 missing)
## stroke < 0.5 to the left, improve=2.121876, (0 missing)
## depression < 0.5 to the left, improve=1.918448, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.612, adj=0.139, (0 split)
## age < 97.5 to the left, agree=0.552, adj=0.005, (0 split)
##
## Node number 434: 238 observations, complexity param=0.0001826494
## predicted class=B1 expected loss=0.5714286 P(node) =0.000866075
## class counts: 102 86 38 9 3
## probabilities: 0.429 0.361 0.160 0.038 0.013
## left son=868 (167 obs) right son=869 (71 obs)
## Primary splits:
## depression < 0.5 to the left, improve=4.834875, (0 missing)
## age < 59.5 to the right, improve=2.461134, (0 missing)
## ihd < 0.5 to the left, improve=1.909944, (0 missing)
## reimbursement2008 < 2285 to the left, improve=1.842456, (0 missing)
## alzheimers < 0.5 to the left, improve=1.113912, (0 missing)
## Surrogate splits:
## age < 97.5 to the left, agree=0.718, adj=0.056, (0 split)
##
## Node number 435: 145 observations
## predicted class=B2 expected loss=0.4689655 P(node) =0.0005276507
## class counts: 39 77 20 8 1
## probabilities: 0.269 0.531 0.138 0.055 0.007
##
## Node number 436: 635 observations
## predicted class=B1 expected loss=0.5023622 P(node) =0.002310746
## class counts: 316 196 93 26 4
## probabilities: 0.498 0.309 0.146 0.041 0.006
##
## Node number 437: 2184 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.595696 P(node) =0.007947511
## class counts: 883 784 346 157 14
## probabilities: 0.404 0.359 0.158 0.072 0.006
## left son=874 (393 obs) right son=875 (1791 obs)
## Primary splits:
## reimbursement2008 < 2315 to the left, improve=4.386891, (0 missing)
## depression < 0.5 to the left, improve=4.376862, (0 missing)
## age < 39.5 to the right, improve=3.004733, (0 missing)
## alzheimers < 0.5 to the left, improve=2.391734, (0 missing)
## stroke < 0.5 to the left, improve=2.171601, (0 missing)
##
## Node number 438: 613 observations, complexity param=8.302246e-05
## predicted class=B1 expected loss=0.6182708 P(node) =0.002230689
## class counts: 234 226 111 36 6
## probabilities: 0.382 0.369 0.181 0.059 0.010
## left son=876 (180 obs) right son=877 (433 obs)
## Primary splits:
## osteoporosis < 0.5 to the right, improve=1.4494640, (0 missing)
## age < 98.5 to the right, improve=1.3979840, (0 missing)
## stroke < 0.5 to the left, improve=0.9190213, (0 missing)
## reimbursement2008 < 2275 to the left, improve=0.8284921, (0 missing)
## depression < 0.5 to the left, improve=0.7804891, (0 missing)
## Surrogate splits:
## reimbursement2008 < 2605 to the right, agree=0.713, adj=0.022, (0 split)
## age < 99.5 to the right, agree=0.711, adj=0.017, (0 split)
##
## Node number 439: 606 observations
## predicted class=B2 expected loss=0.5726073 P(node) =0.002205216
## class counts: 172 259 120 49 6
## probabilities: 0.284 0.427 0.198 0.081 0.010
##
## Node number 440: 143 observations
## predicted class=B1 expected loss=0.3986014 P(node) =0.0005203728
## class counts: 86 37 11 9 0
## probabilities: 0.601 0.259 0.077 0.063 0.000
##
## Node number 441: 374 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5802139 P(node) =0.001360975
## class counts: 157 154 46 16 1
## probabilities: 0.420 0.412 0.123 0.043 0.003
## left son=882 (25 obs) right son=883 (349 obs)
## Primary splits:
## reimbursement2008 < 2315 to the left, improve=4.569334, (0 missing)
## cancer < 0.5 to the left, improve=2.355946, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.361181, (0 missing)
## age < 90.5 to the right, improve=1.103565, (0 missing)
## heart.failure < 0.5 to the left, improve=1.082873, (0 missing)
##
## Node number 442: 18 observations
## predicted class=B1 expected loss=0.2777778 P(node) =6.550147e-05
## class counts: 13 4 0 1 0
## probabilities: 0.722 0.222 0.000 0.056 0.000
##
## Node number 443: 1533 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5942596 P(node) =0.005578542
## class counts: 588 622 223 91 9
## probabilities: 0.384 0.406 0.145 0.059 0.006
## left son=886 (1101 obs) right son=887 (432 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=2.2490350, (0 missing)
## cancer < 0.5 to the left, improve=1.4724050, (0 missing)
## stroke < 0.5 to the left, improve=1.3260620, (0 missing)
## reimbursement2008 < 2435 to the left, improve=1.1404580, (0 missing)
## copd < 0.5 to the left, improve=0.9660973, (0 missing)
##
## Node number 450: 810 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.491358 P(node) =0.002947566
## class counts: 412 257 108 33 0
## probabilities: 0.509 0.317 0.133 0.041 0.000
## left son=900 (117 obs) right son=901 (693 obs)
## Primary splits:
## reimbursement2008 < 11525 to the right, improve=5.1504220, (0 missing)
## bucket2008 < 2.5 to the right, improve=2.8801500, (0 missing)
## age < 34.5 to the left, improve=1.2970260, (0 missing)
## stroke < 0.5 to the left, improve=0.8677994, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5279869, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.928, adj=0.504, (0 split)
##
## Node number 451: 33 observations
## predicted class=B2 expected loss=0.4242424 P(node) =0.000120086
## class counts: 6 19 7 1 0
## probabilities: 0.182 0.576 0.212 0.030 0.000
##
## Node number 452: 3304 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.4757869 P(node) =0.01202316
## class counts: 1732 979 389 183 21
## probabilities: 0.524 0.296 0.118 0.055 0.006
## left son=904 (1626 obs) right son=905 (1678 obs)
## Primary splits:
## reimbursement2008 < 5905 to the right, improve=9.971499, (0 missing)
## bucket2008 < 2.5 to the right, improve=7.851328, (0 missing)
## age < 62.5 to the left, improve=4.441278, (0 missing)
## heart.failure < 0.5 to the left, improve=3.580749, (0 missing)
## kidney < 0.5 to the left, improve=1.765354, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.861, adj=0.717, (0 split)
## kidney < 0.5 to the right, agree=0.637, adj=0.262, (0 split)
## heart.failure < 0.5 to the right, agree=0.608, adj=0.204, (0 split)
## copd < 0.5 to the right, agree=0.590, adj=0.166, (0 split)
## alzheimers < 0.5 to the right, agree=0.558, adj=0.102, (0 split)
##
## Node number 453: 962 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5592516 P(node) =0.00350069
## class counts: 424 364 114 55 5
## probabilities: 0.441 0.378 0.119 0.057 0.005
## left son=906 (857 obs) right son=907 (105 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=3.576264, (0 missing)
## heart.failure < 0.5 to the left, improve=3.114700, (0 missing)
## reimbursement2008 < 59635 to the left, improve=2.145451, (0 missing)
## age < 97.5 to the right, improve=1.742305, (0 missing)
## copd < 0.5 to the left, improve=1.012750, (0 missing)
##
## Node number 454: 1518 observations
## predicted class=B1 expected loss=0.5685112 P(node) =0.005523957
## class counts: 655 520 243 92 8
## probabilities: 0.431 0.343 0.160 0.061 0.005
##
## Node number 455: 706 observations, complexity param=0.0001162314
## predicted class=B2 expected loss=0.6232295 P(node) =0.002569113
## class counts: 233 266 118 77 12
## probabilities: 0.330 0.377 0.167 0.109 0.017
## left son=910 (696 obs) right son=911 (10 obs)
## Primary splits:
## reimbursement2008 < 3155 to the right, improve=3.301177, (0 missing)
## heart.failure < 0.5 to the left, improve=3.232296, (0 missing)
## copd < 0.5 to the left, improve=2.330270, (0 missing)
## alzheimers < 0.5 to the left, improve=1.835216, (0 missing)
## age < 92.5 to the right, improve=1.805094, (0 missing)
##
## Node number 456: 440 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5636364 P(node) =0.001601147
## class counts: 177 192 49 20 2
## probabilities: 0.402 0.436 0.111 0.045 0.005
## left son=912 (58 obs) right son=913 (382 obs)
## Primary splits:
## reimbursement2008 < 3155 to the left, improve=3.3827400, (0 missing)
## age < 80.5 to the right, improve=2.1481460, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6300393, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.5745512, (0 missing)
## depression < 0.5 to the right, improve=0.5547491, (0 missing)
##
## Node number 457: 919 observations
## predicted class=B2 expected loss=0.5505985 P(node) =0.003344214
## class counts: 290 413 154 56 6
## probabilities: 0.316 0.449 0.168 0.061 0.007
##
## Node number 460: 412 observations
## predicted class=B1 expected loss=0.4757282 P(node) =0.001499256
## class counts: 216 120 42 30 4
## probabilities: 0.524 0.291 0.102 0.073 0.010
##
## Node number 461: 302 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6059603 P(node) =0.001098969
## class counts: 102 119 49 31 1
## probabilities: 0.338 0.394 0.162 0.103 0.003
## left son=922 (9 obs) right son=923 (293 obs)
## Primary splits:
## age < 92.5 to the right, improve=2.5766490, (0 missing)
## stroke < 0.5 to the right, improve=1.8961920, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.0608030, (0 missing)
## reimbursement2008 < 32980 to the right, improve=1.0319510, (0 missing)
## kidney < 0.5 to the left, improve=0.7951977, (0 missing)
##
## Node number 462: 282 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.6631206 P(node) =0.00102619
## class counts: 88 95 71 23 5
## probabilities: 0.312 0.337 0.252 0.082 0.018
## left son=924 (220 obs) right son=925 (62 obs)
## Primary splits:
## reimbursement2008 < 27390 to the left, improve=2.933452, (0 missing)
## age < 79.5 to the left, improve=2.171675, (0 missing)
## bucket2008 < 4.5 to the right, improve=1.933271, (0 missing)
## kidney < 0.5 to the left, improve=1.142914, (0 missing)
## ihd < 0.5 to the left, improve=1.125301, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.816, adj=0.161, (0 split)
##
## Node number 463: 162 observations
## predicted class=B2 expected loss=0.5246914 P(node) =0.0005895132
## class counts: 30 77 36 15 4
## probabilities: 0.185 0.475 0.222 0.093 0.025
##
## Node number 472: 159 observations, complexity param=6.167383e-05
## predicted class=B1 expected loss=0.591195 P(node) =0.0005785963
## class counts: 65 51 33 8 2
## probabilities: 0.409 0.321 0.208 0.050 0.013
## left son=944 (76 obs) right son=945 (83 obs)
## Primary splits:
## reimbursement2008 < 11995 to the right, improve=3.4294220, (0 missing)
## age < 65 to the left, improve=1.4674530, (0 missing)
## copd < 0.5 to the left, improve=1.0021090, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7722947, (0 missing)
## bucket2008 < 3.5 to the left, improve=0.4091195, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.673, adj=0.316, (0 split)
## alzheimers < 0.5 to the right, agree=0.591, adj=0.145, (0 split)
## copd < 0.5 to the right, agree=0.572, adj=0.105, (0 split)
## heart.failure < 0.5 to the right, agree=0.560, adj=0.079, (0 split)
## age < 85.5 to the right, agree=0.541, adj=0.039, (0 split)
##
## Node number 473: 586 observations
## predicted class=B2 expected loss=0.6075085 P(node) =0.002132437
## class counts: 167 230 117 64 8
## probabilities: 0.285 0.392 0.200 0.109 0.014
##
## Node number 474: 237 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5738397 P(node) =0.000862436
## class counts: 35 101 72 26 3
## probabilities: 0.148 0.426 0.304 0.110 0.013
## left son=948 (126 obs) right son=949 (111 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.2728500, (0 missing)
## reimbursement2008 < 18275 to the left, improve=1.9530690, (0 missing)
## bucket2008 < 3.5 to the left, improve=1.1622440, (0 missing)
## age < 75.5 to the right, improve=1.0471110, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8993424, (0 missing)
## Surrogate splits:
## age < 86.5 to the left, agree=0.599, adj=0.144, (0 split)
## heart.failure < 0.5 to the left, agree=0.586, adj=0.117, (0 split)
## kidney < 0.5 to the left, agree=0.582, adj=0.108, (0 split)
## depression < 0.5 to the left, agree=0.570, adj=0.081, (0 split)
## stroke < 0.5 to the left, agree=0.565, adj=0.072, (0 split)
##
## Node number 475: 72 observations
## predicted class=B3 expected loss=0.6111111 P(node) =0.0002620059
## class counts: 14 16 28 11 3
## probabilities: 0.194 0.222 0.389 0.153 0.042
##
## Node number 478: 226 observations
## predicted class=B2 expected loss=0.6238938 P(node) =0.0008224073
## class counts: 40 85 74 26 1
## probabilities: 0.177 0.376 0.327 0.115 0.004
##
## Node number 479: 71 observations
## predicted class=B3 expected loss=0.5492958 P(node) =0.0002583669
## class counts: 12 19 32 7 1
## probabilities: 0.169 0.268 0.451 0.099 0.014
##
## Node number 480: 403 observations
## predicted class=B1 expected loss=0.4243176 P(node) =0.001466505
## class counts: 232 86 61 20 4
## probabilities: 0.576 0.213 0.151 0.050 0.010
##
## Node number 481: 2214 observations, complexity param=0.0001439056
## predicted class=B1 expected loss=0.5442638 P(node) =0.008056681
## class counts: 1009 798 290 107 10
## probabilities: 0.456 0.360 0.131 0.048 0.005
## left son=962 (1636 obs) right son=963 (578 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=6.829533, (0 missing)
## depression < 0.5 to the left, improve=5.940363, (0 missing)
## copd < 0.5 to the left, improve=3.544519, (0 missing)
## alzheimers < 0.5 to the left, improve=2.874488, (0 missing)
## age < 57.5 to the left, improve=2.182371, (0 missing)
## Surrogate splits:
## reimbursement2008 < 9185 to the left, agree=0.742, adj=0.01, (0 split)
##
## Node number 482: 5563 observations, complexity param=0.000785946
## predicted class=B1 expected loss=0.6002157 P(node) =0.02024359
## class counts: 2224 2125 853 328 33
## probabilities: 0.400 0.382 0.153 0.059 0.006
## left son=964 (1363 obs) right son=965 (4200 obs)
## Primary splits:
## reimbursement2008 < 8955 to the right, improve=10.881900, (0 missing)
## heart.failure < 0.5 to the left, improve= 9.391652, (0 missing)
## copd < 0.5 to the left, improve= 8.088751, (0 missing)
## bucket2008 < 2.5 to the right, improve= 7.479912, (0 missing)
## age < 31.5 to the left, improve= 2.090877, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.959, adj=0.834, (0 split)
## age < 28.5 to the left, agree=0.755, adj=0.001, (0 split)
##
## Node number 483: 4392 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5778689 P(node) =0.01598236
## class counts: 1379 1854 796 336 27
## probabilities: 0.314 0.422 0.181 0.077 0.006
## left son=966 (2928 obs) right son=967 (1464 obs)
## Primary splits:
## reimbursement2008 < 8325 to the left, improve=7.509791, (0 missing)
## copd < 0.5 to the left, improve=6.714633, (0 missing)
## heart.failure < 0.5 to the left, improve=5.287260, (0 missing)
## bucket2008 < 2.5 to the left, improve=5.126640, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.940785, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.979, adj=0.936, (0 split)
##
## Node number 488: 1063 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5983067 P(node) =0.003868226
## class counts: 336 427 192 95 13
## probabilities: 0.316 0.402 0.181 0.089 0.012
## left son=976 (102 obs) right son=977 (961 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=6.866573, (0 missing)
## heart.failure < 0.5 to the left, improve=4.371067, (0 missing)
## copd < 0.5 to the left, improve=3.836869, (0 missing)
## reimbursement2008 < 18130 to the left, improve=3.632764, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.933358, (0 missing)
##
## Node number 489: 3242 observations
## predicted class=B2 expected loss=0.5009254 P(node) =0.01179754
## class counts: 813 1618 554 233 24
## probabilities: 0.251 0.499 0.171 0.072 0.007
##
## Node number 496: 964 observations, complexity param=0.0001411382
## predicted class=B1 expected loss=0.6307054 P(node) =0.003507968
## class counts: 356 348 167 82 11
## probabilities: 0.369 0.361 0.173 0.085 0.011
## left son=992 (572 obs) right son=993 (392 obs)
## Primary splits:
## depression < 0.5 to the left, improve=5.116701, (0 missing)
## reimbursement2008 < 8255 to the left, improve=4.417859, (0 missing)
## bucket2008 < 2.5 to the left, improve=3.940989, (0 missing)
## heart.failure < 0.5 to the left, improve=3.162605, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.038248, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.601, adj=0.018, (0 split)
## reimbursement2008 < 14855 to the left, agree=0.598, adj=0.010, (0 split)
## copd < 0.5 to the left, agree=0.594, adj=0.003, (0 split)
##
## Node number 497: 6822 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6002639 P(node) =0.02482506
## class counts: 1626 2727 1500 847 122
## probabilities: 0.238 0.400 0.220 0.124 0.018
## left son=994 (3172 obs) right son=995 (3650 obs)
## Primary splits:
## reimbursement2008 < 6325 to the left, improve=11.481700, (0 missing)
## depression < 0.5 to the left, improve=11.166950, (0 missing)
## bucket2008 < 2.5 to the left, improve= 7.602059, (0 missing)
## osteoporosis < 0.5 to the left, improve= 6.404955, (0 missing)
## copd < 0.5 to the left, improve= 5.945024, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.869, adj=0.717, (0 split)
## copd < 0.5 to the left, agree=0.576, adj=0.088, (0 split)
## heart.failure < 0.5 to the left, agree=0.570, adj=0.076, (0 split)
## alzheimers < 0.5 to the left, agree=0.545, adj=0.020, (0 split)
## age < 31.5 to the left, agree=0.536, adj=0.003, (0 split)
##
## Node number 504: 1291 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5530596 P(node) =0.004697911
## class counts: 183 577 280 219 32
## probabilities: 0.142 0.447 0.217 0.170 0.025
## left son=1008 (973 obs) right son=1009 (318 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=5.737334, (0 missing)
## reimbursement2008 < 32055 to the left, improve=1.892346, (0 missing)
## age < 84.5 to the right, improve=1.763761, (0 missing)
## ihd < 0.5 to the left, improve=1.710826, (0 missing)
## alzheimers < 0.5 to the right, improve=1.100420, (0 missing)
## Surrogate splits:
## age < 28.5 to the right, agree=0.755, adj=0.006, (0 split)
##
## Node number 505: 2054 observations
## predicted class=B2 expected loss=0.5978578 P(node) =0.007474445
## class counts: 189 826 557 413 69
## probabilities: 0.092 0.402 0.271 0.201 0.034
##
## Node number 506: 520 observations
## predicted class=B2 expected loss=0.5769231 P(node) =0.001892265
## class counts: 50 220 120 108 22
## probabilities: 0.096 0.423 0.231 0.208 0.042
##
## Node number 507: 1537 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6363045 P(node) =0.005593098
## class counts: 87 559 353 446 92
## probabilities: 0.057 0.364 0.230 0.290 0.060
## left son=1014 (1286 obs) right son=1015 (251 obs)
## Primary splits:
## age < 62.5 to the right, improve=4.292573, (0 missing)
## reimbursement2008 < 43950 to the left, improve=3.358938, (0 missing)
## cancer < 0.5 to the right, improve=2.803709, (0 missing)
## heart.failure < 0.5 to the left, improve=1.956332, (0 missing)
## stroke < 0.5 to the left, improve=1.605851, (0 missing)
##
## Node number 508: 2489 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.7219767 P(node) =0.009057397
## class counts: 541 692 436 670 150
## probabilities: 0.217 0.278 0.175 0.269 0.060
## left son=1016 (1317 obs) right son=1017 (1172 obs)
## Primary splits:
## copd < 0.5 to the right, improve=9.293163, (0 missing)
## reimbursement2008 < 23175 to the left, improve=7.265866, (0 missing)
## ihd < 0.5 to the left, improve=7.177016, (0 missing)
## heart.failure < 0.5 to the left, improve=4.187307, (0 missing)
## bucket2008 < 4.5 to the left, improve=3.684093, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the right, agree=0.575, adj=0.097, (0 split)
## alzheimers < 0.5 to the right, agree=0.556, adj=0.057, (0 split)
## ihd < 0.5 to the right, agree=0.555, adj=0.055, (0 split)
## reimbursement2008 < 27380 to the right, agree=0.544, adj=0.032, (0 split)
## age < 52.5 to the right, agree=0.532, adj=0.005, (0 split)
##
## Node number 509: 2809 observations
## predicted class=B2 expected loss=0.6600214 P(node) =0.01022187
## class counts: 367 955 615 694 178
## probabilities: 0.131 0.340 0.219 0.247 0.063
##
## Node number 812: 95 observations
## predicted class=B1 expected loss=0.4315789 P(node) =0.0003457022
## class counts: 54 25 11 5 0
## probabilities: 0.568 0.263 0.116 0.053 0.000
##
## Node number 813: 33 observations
## predicted class=B2 expected loss=0.4848485 P(node) =0.000120086
## class counts: 10 17 3 3 0
## probabilities: 0.303 0.515 0.091 0.091 0.000
##
## Node number 826: 596 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5536913 P(node) =0.002168826
## class counts: 266 214 77 34 5
## probabilities: 0.446 0.359 0.129 0.057 0.008
## left son=1652 (555 obs) right son=1653 (41 obs)
## Primary splits:
## reimbursement2008 < 1765 to the right, improve=2.4482060, (0 missing)
## age < 81.5 to the right, improve=2.3548750, (0 missing)
## ihd < 0.5 to the left, improve=2.3213280, (0 missing)
## cancer < 0.5 to the left, improve=2.1512770, (0 missing)
## depression < 0.5 to the left, improve=0.5755867, (0 missing)
##
## Node number 827: 38 observations
## predicted class=B2 expected loss=0.5526316 P(node) =0.0001382809
## class counts: 7 17 6 8 0
## probabilities: 0.184 0.447 0.158 0.211 0.000
##
## Node number 846: 22 observations
## predicted class=B1 expected loss=0.4545455 P(node) =8.005735e-05
## class counts: 12 5 4 1 0
## probabilities: 0.545 0.227 0.182 0.045 0.000
##
## Node number 847: 96 observations
## predicted class=B2 expected loss=0.4791667 P(node) =0.0003493412
## class counts: 29 50 12 5 0
## probabilities: 0.302 0.521 0.125 0.052 0.000
##
## Node number 848: 385 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5454545 P(node) =0.001401004
## class counts: 175 146 39 23 2
## probabilities: 0.455 0.379 0.101 0.060 0.005
## left son=1696 (263 obs) right son=1697 (122 obs)
## Primary splits:
## age < 80.5 to the left, improve=2.5496070, (0 missing)
## ihd < 0.5 to the left, improve=1.7465940, (0 missing)
## reimbursement2008 < 3025 to the right, improve=0.9395484, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7038989, (0 missing)
## depression < 0.5 to the left, improve=0.3133237, (0 missing)
##
## Node number 849: 110 observations
## predicted class=B2 expected loss=0.4909091 P(node) =0.0004002868
## class counts: 40 56 11 3 0
## probabilities: 0.364 0.509 0.100 0.027 0.000
##
## Node number 866: 1499 observations
## predicted class=B1 expected loss=0.490994 P(node) =0.005454817
## class counts: 763 492 189 52 3
## probabilities: 0.509 0.328 0.126 0.035 0.002
##
## Node number 867: 1228 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5537459 P(node) =0.004468656
## class counts: 548 456 161 57 6
## probabilities: 0.446 0.371 0.131 0.046 0.005
## left son=1734 (171 obs) right son=1735 (1057 obs)
## Primary splits:
## reimbursement2008 < 2995 to the right, improve=3.399761, (0 missing)
## bucket2008 < 1.5 to the right, improve=3.399761, (0 missing)
## age < 83.5 to the left, improve=2.697325, (0 missing)
## kidney < 0.5 to the left, improve=2.465389, (0 missing)
## depression < 0.5 to the left, improve=1.722272, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=1, adj=1, (0 split)
##
## Node number 868: 167 observations
## predicted class=B1 expected loss=0.502994 P(node) =0.0006077081
## class counts: 83 50 25 7 2
## probabilities: 0.497 0.299 0.150 0.042 0.012
##
## Node number 869: 71 observations
## predicted class=B2 expected loss=0.4929577 P(node) =0.0002583669
## class counts: 19 36 13 2 1
## probabilities: 0.268 0.507 0.183 0.028 0.014
##
## Node number 874: 393 observations
## predicted class=B1 expected loss=0.5139949 P(node) =0.001430115
## class counts: 191 134 46 20 2
## probabilities: 0.486 0.341 0.117 0.051 0.005
##
## Node number 875: 1791 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.6136237 P(node) =0.006517396
## class counts: 692 650 300 137 12
## probabilities: 0.386 0.363 0.168 0.076 0.007
## left son=1750 (1752 obs) right son=1751 (39 obs)
## Primary splits:
## age < 39.5 to the right, improve=3.631907, (0 missing)
## depression < 0.5 to the left, improve=3.598994, (0 missing)
## reimbursement2008 < 2475 to the left, improve=1.828499, (0 missing)
## alzheimers < 0.5 to the left, improve=1.790378, (0 missing)
## stroke < 0.5 to the left, improve=1.609073, (0 missing)
##
## Node number 876: 180 observations, complexity param=8.302246e-05
## predicted class=B1 expected loss=0.5666667 P(node) =0.0006550147
## class counts: 78 68 23 11 0
## probabilities: 0.433 0.378 0.128 0.061 0.000
## left son=1752 (112 obs) right son=1753 (68 obs)
## Primary splits:
## reimbursement2008 < 2455 to the left, improve=3.4787820, (0 missing)
## age < 88.5 to the left, improve=1.3486290, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8074074, (0 missing)
## copd < 0.5 to the left, improve=0.6952328, (0 missing)
## ihd < 0.5 to the right, improve=0.3305556, (0 missing)
##
## Node number 877: 433 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.6351039 P(node) =0.001575674
## class counts: 156 158 88 25 6
## probabilities: 0.360 0.365 0.203 0.058 0.014
## left son=1754 (403 obs) right son=1755 (30 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=1.8988510, (0 missing)
## age < 45.5 to the left, improve=1.3010460, (0 missing)
## reimbursement2008 < 2255 to the left, improve=1.2799960, (0 missing)
## depression < 0.5 to the left, improve=1.2338070, (0 missing)
## cancer < 0.5 to the right, improve=0.6525541, (0 missing)
##
## Node number 882: 25 observations
## predicted class=B2 expected loss=0.24 P(node) =9.097426e-05
## class counts: 6 19 0 0 0
## probabilities: 0.240 0.760 0.000 0.000 0.000
##
## Node number 883: 349 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5673352 P(node) =0.001270001
## class counts: 151 135 46 16 1
## probabilities: 0.433 0.387 0.132 0.046 0.003
## left son=1766 (336 obs) right son=1767 (13 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.2574550, (0 missing)
## age < 69.5 to the left, improve=1.5047970, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.0520090, (0 missing)
## heart.failure < 0.5 to the left, improve=0.8771249, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.8193194, (0 missing)
##
## Node number 886: 1101 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.595822 P(node) =0.004006506
## class counts: 445 444 150 57 5
## probabilities: 0.404 0.403 0.136 0.052 0.005
## left son=1772 (1057 obs) right son=1773 (44 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=2.0669290, (0 missing)
## age < 46.5 to the left, improve=1.0570040, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.9632398, (0 missing)
## cancer < 0.5 to the left, improve=0.9197684, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7998904, (0 missing)
##
## Node number 887: 432 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.587963 P(node) =0.001572035
## class counts: 143 178 73 34 4
## probabilities: 0.331 0.412 0.169 0.079 0.009
## left son=1774 (403 obs) right son=1775 (29 obs)
## Primary splits:
## reimbursement2008 < 2215 to the right, improve=1.979831, (0 missing)
## depression < 0.5 to the left, improve=1.399095, (0 missing)
## age < 65.5 to the right, improve=1.336452, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.190308, (0 missing)
## heart.failure < 0.5 to the left, improve=1.062301, (0 missing)
##
## Node number 900: 117 observations
## predicted class=B1 expected loss=0.3418803 P(node) =0.0004257595
## class counts: 77 25 8 7 0
## probabilities: 0.658 0.214 0.068 0.060 0.000
##
## Node number 901: 693 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5165945 P(node) =0.002521807
## class counts: 335 232 100 26 0
## probabilities: 0.483 0.335 0.144 0.038 0.000
## left son=1802 (684 obs) right son=1803 (9 obs)
## Primary splits:
## reimbursement2008 < 11105 to the left, improve=2.2165640, (0 missing)
## alzheimers < 0.5 to the left, improve=1.2536740, (0 missing)
## stroke < 0.5 to the left, improve=0.8354877, (0 missing)
## age < 34 to the left, improve=0.8168384, (0 missing)
## copd < 0.5 to the left, improve=0.3325841, (0 missing)
##
## Node number 904: 1626 observations
## predicted class=B1 expected loss=0.4391144 P(node) =0.005916966
## class counts: 912 414 190 99 11
## probabilities: 0.561 0.255 0.117 0.061 0.007
##
## Node number 905: 1678 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.511323 P(node) =0.006106192
## class counts: 820 565 199 84 10
## probabilities: 0.489 0.337 0.119 0.050 0.006
## left son=1810 (1608 obs) right son=1811 (70 obs)
## Primary splits:
## reimbursement2008 < 5695 to the left, improve=3.4228850, (0 missing)
## age < 54.5 to the left, improve=3.4011680, (0 missing)
## kidney < 0.5 to the left, improve=2.7930360, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6256038, (0 missing)
## stroke < 0.5 to the left, improve=0.4872841, (0 missing)
##
## Node number 906: 857 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5425904 P(node) =0.003118598
## class counts: 392 313 100 48 4
## probabilities: 0.457 0.365 0.117 0.056 0.005
## left son=1812 (405 obs) right son=1813 (452 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=1.8446560, (0 missing)
## reimbursement2008 < 8165 to the right, improve=1.5511950, (0 missing)
## age < 58.5 to the right, improve=1.3750490, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.0178390, (0 missing)
## kidney < 0.5 to the left, improve=0.9078092, (0 missing)
## Surrogate splits:
## reimbursement2008 < 7010 to the left, agree=0.629, adj=0.215, (0 split)
## bucket2008 < 2.5 to the left, agree=0.611, adj=0.178, (0 split)
## kidney < 0.5 to the left, agree=0.585, adj=0.121, (0 split)
## copd < 0.5 to the left, agree=0.583, adj=0.119, (0 split)
## age < 77.5 to the left, agree=0.555, adj=0.059, (0 split)
##
## Node number 907: 105 observations
## predicted class=B2 expected loss=0.5142857 P(node) =0.0003820919
## class counts: 32 51 14 7 1
## probabilities: 0.305 0.486 0.133 0.067 0.010
##
## Node number 910: 696 observations, complexity param=0.0001162314
## predicted class=B2 expected loss=0.6192529 P(node) =0.002532723
## class counts: 232 265 112 75 12
## probabilities: 0.333 0.381 0.161 0.108 0.017
## left son=1820 (177 obs) right son=1821 (519 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.566495, (0 missing)
## copd < 0.5 to the left, improve=2.688595, (0 missing)
## age < 70.5 to the right, improve=1.873250, (0 missing)
## alzheimers < 0.5 to the left, improve=1.737201, (0 missing)
## reimbursement2008 < 7405 to the left, improve=1.699902, (0 missing)
##
## Node number 911: 10 observations
## predicted class=B3 expected loss=0.4 P(node) =3.63897e-05
## class counts: 1 1 6 2 0
## probabilities: 0.100 0.100 0.600 0.200 0.000
##
## Node number 912: 58 observations
## predicted class=B2 expected loss=0.3793103 P(node) =0.0002110603
## class counts: 20 36 0 1 1
## probabilities: 0.345 0.621 0.000 0.017 0.017
##
## Node number 913: 382 observations, complexity param=5.313437e-05
## predicted class=B1 expected loss=0.5890052 P(node) =0.001390087
## class counts: 157 156 49 19 1
## probabilities: 0.411 0.408 0.128 0.050 0.003
## left son=1826 (25 obs) right son=1827 (357 obs)
## Primary splits:
## reimbursement2008 < 3245 to the left, improve=2.6514540, (0 missing)
## age < 80.5 to the right, improve=2.1655330, (0 missing)
## depression < 0.5 to the left, improve=1.2095670, (0 missing)
## kidney < 0.5 to the left, improve=1.1024610, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7876318, (0 missing)
##
## Node number 922: 9 observations
## predicted class=B1 expected loss=0.3333333 P(node) =3.275073e-05
## class counts: 6 0 1 1 1
## probabilities: 0.667 0.000 0.111 0.111 0.111
##
## Node number 923: 293 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5938567 P(node) =0.001066218
## class counts: 96 119 48 30 0
## probabilities: 0.328 0.406 0.164 0.102 0.000
## left son=1846 (39 obs) right son=1847 (254 obs)
## Primary splits:
## stroke < 0.5 to the right, improve=2.1766940, (0 missing)
## age < 65.5 to the right, improve=1.5636490, (0 missing)
## reimbursement2008 < 11660 to the right, improve=1.0372370, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9901623, (0 missing)
## kidney < 0.5 to the left, improve=0.7188410, (0 missing)
## Surrogate splits:
## reimbursement2008 < 79760 to the right, agree=0.87, adj=0.026, (0 split)
##
## Node number 924: 220 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.65 P(node) =0.0008005735
## class counts: 77 66 57 16 4
## probabilities: 0.350 0.300 0.259 0.073 0.018
## left son=1848 (132 obs) right son=1849 (88 obs)
## Primary splits:
## reimbursement2008 < 12810 to the right, improve=2.5787880, (0 missing)
## age < 76.5 to the left, improve=2.1718510, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.0384950, (0 missing)
## stroke < 0.5 to the left, improve=0.8392769, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7969697, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.636, adj=0.091, (0 split)
##
## Node number 925: 62 observations
## predicted class=B2 expected loss=0.5322581 P(node) =0.0002256162
## class counts: 11 29 14 7 1
## probabilities: 0.177 0.468 0.226 0.113 0.016
##
## Node number 944: 76 observations
## predicted class=B1 expected loss=0.4736842 P(node) =0.0002765618
## class counts: 40 18 12 5 1
## probabilities: 0.526 0.237 0.158 0.066 0.013
##
## Node number 945: 83 observations
## predicted class=B2 expected loss=0.6024096 P(node) =0.0003020345
## class counts: 25 33 21 3 1
## probabilities: 0.301 0.398 0.253 0.036 0.012
##
## Node number 948: 126 observations
## predicted class=B2 expected loss=0.5079365 P(node) =0.0004585103
## class counts: 22 62 33 9 0
## probabilities: 0.175 0.492 0.262 0.071 0.000
##
## Node number 949: 111 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6486486 P(node) =0.0004039257
## class counts: 13 39 39 17 3
## probabilities: 0.117 0.351 0.351 0.153 0.027
## left son=1898 (54 obs) right son=1899 (57 obs)
## Primary splits:
## age < 75.5 to the left, improve=1.9702330, (0 missing)
## reimbursement2008 < 23940 to the left, improve=1.5183950, (0 missing)
## stroke < 0.5 to the right, improve=1.4096100, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.1218360, (0 missing)
## kidney < 0.5 to the right, improve=0.9749865, (0 missing)
## Surrogate splits:
## reimbursement2008 < 36125 to the right, agree=0.586, adj=0.148, (0 split)
## depression < 0.5 to the right, agree=0.568, adj=0.111, (0 split)
## alzheimers < 0.5 to the right, agree=0.550, adj=0.074, (0 split)
## bucket2008 < 4.5 to the right, agree=0.532, adj=0.037, (0 split)
##
## Node number 962: 1636 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5177262 P(node) =0.005953356
## class counts: 789 562 198 80 7
## probabilities: 0.482 0.344 0.121 0.049 0.004
## left son=1924 (1127 obs) right son=1925 (509 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.275005, (0 missing)
## depression < 0.5 to the left, improve=2.200834, (0 missing)
## age < 56.5 to the right, improve=2.161392, (0 missing)
## reimbursement2008 < 3635 to the left, improve=1.571205, (0 missing)
## copd < 0.5 to the left, improve=1.483908, (0 missing)
## Surrogate splits:
## reimbursement2008 < 9210 to the left, agree=0.69, adj=0.004, (0 split)
##
## Node number 963: 578 observations, complexity param=0.0001439056
## predicted class=B2 expected loss=0.5916955 P(node) =0.002103325
## class counts: 220 236 92 27 3
## probabilities: 0.381 0.408 0.159 0.047 0.005
## left son=1926 (339 obs) right son=1927 (239 obs)
## Primary splits:
## depression < 0.5 to the left, improve=4.643194, (0 missing)
## copd < 0.5 to the left, improve=3.299852, (0 missing)
## reimbursement2008 < 4535 to the left, improve=1.771265, (0 missing)
## alzheimers < 0.5 to the left, improve=1.392171, (0 missing)
## age < 39.5 to the right, improve=1.271983, (0 missing)
## Surrogate splits:
## age < 43.5 to the right, agree=0.602, adj=0.038, (0 split)
##
## Node number 964: 1363 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5561262 P(node) =0.004959917
## class counts: 605 435 219 92 12
## probabilities: 0.444 0.319 0.161 0.067 0.009
## left son=1928 (798 obs) right son=1929 (565 obs)
## Primary splits:
## copd < 0.5 to the left, improve=7.963265, (0 missing)
## heart.failure < 0.5 to the left, improve=4.055475, (0 missing)
## reimbursement2008 < 29005 to the left, improve=3.506334, (0 missing)
## stroke < 0.5 to the left, improve=1.818093, (0 missing)
## alzheimers < 0.5 to the left, improve=1.815648, (0 missing)
## Surrogate splits:
## reimbursement2008 < 55265 to the left, agree=0.589, adj=0.009, (0 split)
## bucket2008 < 4.5 to the left, agree=0.589, adj=0.009, (0 split)
## age < 27.5 to the right, agree=0.588, adj=0.005, (0 split)
##
## Node number 965: 4200 observations, complexity param=0.0004649258
## predicted class=B2 expected loss=0.597619 P(node) =0.01528368
## class counts: 1619 1690 634 236 21
## probabilities: 0.385 0.402 0.151 0.056 0.005
## left son=1930 (1953 obs) right son=1931 (2247 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=7.153470, (0 missing)
## copd < 0.5 to the left, improve=3.796544, (0 missing)
## age < 49.5 to the left, improve=2.763159, (0 missing)
## reimbursement2008 < 3415 to the left, improve=2.562311, (0 missing)
## alzheimers < 0.5 to the left, improve=1.867356, (0 missing)
## Surrogate splits:
## reimbursement2008 < 4495 to the left, agree=0.563, adj=0.060, (0 split)
## copd < 0.5 to the left, agree=0.551, adj=0.035, (0 split)
## alzheimers < 0.5 to the left, agree=0.540, adj=0.010, (0 split)
## age < 45.5 to the left, agree=0.536, adj=0.002, (0 split)
##
## Node number 966: 2928 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5529372 P(node) =0.01065491
## class counts: 904 1309 506 190 19
## probabilities: 0.309 0.447 0.173 0.065 0.006
## left son=1932 (1987 obs) right son=1933 (941 obs)
## Primary splits:
## copd < 0.5 to the left, improve=5.088653, (0 missing)
## osteoporosis < 0.5 to the left, improve=4.205973, (0 missing)
## reimbursement2008 < 8045 to the left, improve=3.909055, (0 missing)
## heart.failure < 0.5 to the left, improve=3.356901, (0 missing)
## stroke < 0.5 to the left, improve=3.071040, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8235 to the left, agree=0.68, adj=0.005, (0 split)
##
## Node number 967: 1464 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6277322 P(node) =0.005327453
## class counts: 475 545 290 146 8
## probabilities: 0.324 0.372 0.198 0.100 0.005
## left son=1934 (36 obs) right son=1935 (1428 obs)
## Primary splits:
## reimbursement2008 < 8485 to the left, improve=3.191750, (0 missing)
## age < 78.5 to the left, improve=2.281932, (0 missing)
## heart.failure < 0.5 to the left, improve=2.180745, (0 missing)
## stroke < 0.5 to the left, improve=1.944689, (0 missing)
## copd < 0.5 to the left, improve=1.512341, (0 missing)
##
## Node number 976: 102 observations
## predicted class=B1 expected loss=0.4803922 P(node) =0.000371175
## class counts: 53 28 13 7 1
## probabilities: 0.520 0.275 0.127 0.069 0.010
##
## Node number 977: 961 observations
## predicted class=B2 expected loss=0.5848075 P(node) =0.003497051
## class counts: 283 399 179 88 12
## probabilities: 0.294 0.415 0.186 0.092 0.012
##
## Node number 992: 572 observations, complexity param=0.0001411382
## predicted class=B1 expected loss=0.5909091 P(node) =0.002081491
## class counts: 234 183 92 57 6
## probabilities: 0.409 0.320 0.161 0.100 0.010
## left son=1984 (101 obs) right son=1985 (471 obs)
## Primary splits:
## reimbursement2008 < 3545 to the left, improve=4.251847, (0 missing)
## bucket2008 < 2.5 to the left, improve=2.618377, (0 missing)
## age < 69.5 to the right, improve=2.566628, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.664728, (0 missing)
## heart.failure < 0.5 to the left, improve=1.532473, (0 missing)
##
## Node number 993: 392 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5790816 P(node) =0.001426476
## class counts: 122 165 75 25 5
## probabilities: 0.311 0.421 0.191 0.064 0.013
## left son=1986 (9 obs) right son=1987 (383 obs)
## Primary splits:
## reimbursement2008 < 14460 to the right, improve=2.744913, (0 missing)
## age < 48.5 to the left, improve=1.556717, (0 missing)
## copd < 0.5 to the left, improve=1.522824, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.319075, (0 missing)
## heart.failure < 0.5 to the left, improve=1.292043, (0 missing)
##
## Node number 994: 3172 observations
## predicted class=B2 expected loss=0.5630517 P(node) =0.01154281
## class counts: 709 1386 688 337 52
## probabilities: 0.224 0.437 0.217 0.106 0.016
##
## Node number 995: 3650 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6326027 P(node) =0.01328224
## class counts: 917 1341 812 510 70
## probabilities: 0.251 0.367 0.222 0.140 0.019
## left son=1990 (2424 obs) right son=1991 (1226 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=8.187206, (0 missing)
## depression < 0.5 to the left, improve=5.387571, (0 missing)
## copd < 0.5 to the left, improve=5.189054, (0 missing)
## alzheimers < 0.5 to the left, improve=3.710849, (0 missing)
## heart.failure < 0.5 to the left, improve=2.971629, (0 missing)
##
## Node number 1008: 973 observations
## predicted class=B2 expected loss=0.5611511 P(node) =0.003540718
## class counts: 160 427 184 174 28
## probabilities: 0.164 0.439 0.189 0.179 0.029
##
## Node number 1009: 318 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5283019 P(node) =0.001157193
## class counts: 23 150 96 45 4
## probabilities: 0.072 0.472 0.302 0.142 0.013
## left son=2018 (293 obs) right son=2019 (25 obs)
## Primary splits:
## reimbursement2008 < 16525 to the right, improve=7.797134, (0 missing)
## ihd < 0.5 to the left, improve=3.194748, (0 missing)
## heart.failure < 0.5 to the left, improve=1.170376, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.159119, (0 missing)
## age < 55 to the right, improve=1.113448, (0 missing)
##
## Node number 1014: 1286 observations
## predicted class=B2 expected loss=0.6251944 P(node) =0.004679716
## class counts: 74 482 303 348 79
## probabilities: 0.058 0.375 0.236 0.271 0.061
##
## Node number 1015: 251 observations, complexity param=6.088314e-05
## predicted class=B4 expected loss=0.6095618 P(node) =0.0009133816
## class counts: 13 77 50 98 13
## probabilities: 0.052 0.307 0.199 0.390 0.052
## left son=2030 (237 obs) right son=2031 (14 obs)
## Primary splits:
## reimbursement2008 < 101585 to the left, improve=3.425401, (0 missing)
## age < 61.5 to the left, improve=2.440583, (0 missing)
## heart.failure < 0.5 to the left, improve=2.158559, (0 missing)
## alzheimers < 0.5 to the left, improve=2.020557, (0 missing)
## cancer < 0.5 to the right, improve=1.778561, (0 missing)
##
## Node number 1016: 1317 observations, complexity param=0.0001439056
## predicted class=B2 expected loss=0.6757783 P(node) =0.004792524
## class counts: 269 427 234 313 74
## probabilities: 0.204 0.324 0.178 0.238 0.056
## left son=2032 (72 obs) right son=2033 (1245 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=5.587185, (0 missing)
## reimbursement2008 < 22435 to the left, improve=5.039175, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.150989, (0 missing)
## age < 50.5 to the right, improve=2.709964, (0 missing)
## heart.failure < 0.5 to the left, improve=2.161876, (0 missing)
##
## Node number 1017: 1172 observations, complexity param=0.0002036818
## predicted class=B4 expected loss=0.6953925 P(node) =0.004264873
## class counts: 272 265 202 357 76
## probabilities: 0.232 0.226 0.172 0.305 0.065
## left son=2034 (191 obs) right son=2035 (981 obs)
## Primary splits:
## reimbursement2008 < 43640 to the right, improve=6.105110, (0 missing)
## bucket2008 < 4.5 to the right, improve=4.295055, (0 missing)
## ihd < 0.5 to the left, improve=2.740224, (0 missing)
## heart.failure < 0.5 to the left, improve=2.395917, (0 missing)
## alzheimers < 0.5 to the right, improve=1.864237, (0 missing)
## Surrogate splits:
## bucket2008 < 4.5 to the right, agree=0.925, adj=0.539, (0 split)
##
## Node number 1652: 555 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5441441 P(node) =0.002019629
## class counts: 253 193 76 29 4
## probabilities: 0.456 0.348 0.137 0.052 0.007
## left son=3304 (265 obs) right son=3305 (290 obs)
## Primary splits:
## reimbursement2008 < 1955 to the left, improve=2.2492970, (0 missing)
## age < 44.5 to the right, improve=2.2010530, (0 missing)
## cancer < 0.5 to the left, improve=2.0271900, (0 missing)
## ihd < 0.5 to the left, improve=2.0227740, (0 missing)
## stroke < 0.5 to the left, improve=0.6965966, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.539, adj=0.034, (0 split)
## age < 90.5 to the right, agree=0.532, adj=0.019, (0 split)
## cancer < 0.5 to the right, agree=0.528, adj=0.011, (0 split)
##
## Node number 1653: 41 observations
## predicted class=B2 expected loss=0.4878049 P(node) =0.0001491978
## class counts: 13 21 1 5 1
## probabilities: 0.317 0.512 0.024 0.122 0.024
##
## Node number 1696: 263 observations
## predicted class=B1 expected loss=0.4980989 P(node) =0.0009570492
## class counts: 132 94 26 11 0
## probabilities: 0.502 0.357 0.099 0.042 0.000
##
## Node number 1697: 122 observations
## predicted class=B2 expected loss=0.5737705 P(node) =0.0004439544
## class counts: 43 52 13 12 2
## probabilities: 0.352 0.426 0.107 0.098 0.016
##
## Node number 1734: 171 observations
## predicted class=B1 expected loss=0.4736842 P(node) =0.0006222639
## class counts: 90 46 24 10 1
## probabilities: 0.526 0.269 0.140 0.058 0.006
##
## Node number 1735: 1057 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5666982 P(node) =0.003846392
## class counts: 458 410 137 47 5
## probabilities: 0.433 0.388 0.130 0.044 0.005
## left son=3470 (840 obs) right son=3471 (217 obs)
## Primary splits:
## age < 83.5 to the left, improve=3.809819, (0 missing)
## kidney < 0.5 to the left, improve=2.564065, (0 missing)
## depression < 0.5 to the left, improve=1.351420, (0 missing)
## copd < 0.5 to the left, improve=1.145117, (0 missing)
## reimbursement2008 < 2975 to the left, improve=1.026292, (0 missing)
##
## Node number 1750: 1752 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.6084475 P(node) =0.006375476
## class counts: 686 629 294 131 12
## probabilities: 0.392 0.359 0.168 0.075 0.007
## left son=3500 (1099 obs) right son=3501 (653 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.008256, (0 missing)
## age < 97.5 to the left, improve=2.167182, (0 missing)
## reimbursement2008 < 3055 to the left, improve=1.739250, (0 missing)
## alzheimers < 0.5 to the left, improve=1.536121, (0 missing)
## stroke < 0.5 to the left, improve=1.306511, (0 missing)
## Surrogate splits:
## age < 41.5 to the right, agree=0.628, adj=0.002, (0 split)
##
## Node number 1751: 39 observations
## predicted class=B2 expected loss=0.4615385 P(node) =0.0001419198
## class counts: 6 21 6 6 0
## probabilities: 0.154 0.538 0.154 0.154 0.000
##
## Node number 1752: 112 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0004075647
## class counts: 56 33 14 9 0
## probabilities: 0.500 0.295 0.125 0.080 0.000
##
## Node number 1753: 68 observations
## predicted class=B2 expected loss=0.4852941 P(node) =0.00024745
## class counts: 22 35 9 2 0
## probabilities: 0.324 0.515 0.132 0.029 0.000
##
## Node number 1754: 403 observations, complexity param=6.272808e-05
## predicted class=B1 expected loss=0.6253102 P(node) =0.001466505
## class counts: 151 147 79 20 6
## probabilities: 0.375 0.365 0.196 0.050 0.015
## left son=3508 (382 obs) right son=3509 (21 obs)
## Primary splits:
## reimbursement2008 < 2585 to the left, improve=1.3835870, (0 missing)
## age < 45.5 to the left, improve=1.1912610, (0 missing)
## depression < 0.5 to the left, improve=0.8974996, (0 missing)
## cancer < 0.5 to the right, improve=0.7248908, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2961750, (0 missing)
##
## Node number 1755: 30 observations
## predicted class=B2 expected loss=0.6333333 P(node) =0.0001091691
## class counts: 5 11 9 5 0
## probabilities: 0.167 0.367 0.300 0.167 0.000
##
## Node number 1766: 336 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5535714 P(node) =0.001222694
## class counts: 150 128 44 14 0
## probabilities: 0.446 0.381 0.131 0.042 0.000
## left son=3532 (322 obs) right son=3533 (14 obs)
## Primary splits:
## age < 90.5 to the left, improve=1.4565220, (0 missing)
## heart.failure < 0.5 to the left, improve=1.1063780, (0 missing)
## reimbursement2008 < 2325 to the right, improve=0.8683190, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8476676, (0 missing)
## stroke < 0.5 to the left, improve=0.4098214, (0 missing)
##
## Node number 1767: 13 observations
## predicted class=B2 expected loss=0.4615385 P(node) =4.730662e-05
## class counts: 1 7 2 2 1
## probabilities: 0.077 0.538 0.154 0.154 0.077
##
## Node number 1772: 1057 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.589404 P(node) =0.003846392
## class counts: 434 420 145 54 4
## probabilities: 0.411 0.397 0.137 0.051 0.004
## left son=3544 (1008 obs) right son=3545 (49 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=1.0144070, (0 missing)
## age < 46.5 to the left, improve=1.0088960, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.9481247, (0 missing)
## copd < 0.5 to the left, improve=0.6576908, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5672579, (0 missing)
##
## Node number 1773: 44 observations
## predicted class=B2 expected loss=0.4545455 P(node) =0.0001601147
## class counts: 11 24 5 3 1
## probabilities: 0.250 0.545 0.114 0.068 0.023
##
## Node number 1774: 403 observations
## predicted class=B2 expected loss=0.5756824 P(node) =0.001466505
## class counts: 130 171 70 28 4
## probabilities: 0.323 0.424 0.174 0.069 0.010
##
## Node number 1775: 29 observations
## predicted class=B1 expected loss=0.5517241 P(node) =0.0001055301
## class counts: 13 7 3 6 0
## probabilities: 0.448 0.241 0.103 0.207 0.000
##
## Node number 1802: 684 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5146199 P(node) =0.002489056
## class counts: 332 231 95 26 0
## probabilities: 0.485 0.338 0.139 0.038 0.000
## left son=3604 (286 obs) right son=3605 (398 obs)
## Primary splits:
## reimbursement2008 < 4365 to the left, improve=1.4254390, (0 missing)
## alzheimers < 0.5 to the left, improve=1.1902870, (0 missing)
## stroke < 0.5 to the left, improve=0.8341619, (0 missing)
## age < 34 to the left, improve=0.8175360, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3590913, (0 missing)
##
## Node number 1803: 9 observations
## predicted class=B3 expected loss=0.4444444 P(node) =3.275073e-05
## class counts: 3 1 5 0 0
## probabilities: 0.333 0.111 0.556 0.000 0.000
##
## Node number 1810: 1608 observations
## predicted class=B1 expected loss=0.5062189 P(node) =0.005851465
## class counts: 794 529 193 82 10
## probabilities: 0.494 0.329 0.120 0.051 0.006
##
## Node number 1811: 70 observations
## predicted class=B2 expected loss=0.4857143 P(node) =0.0002547279
## class counts: 26 36 6 2 0
## probabilities: 0.371 0.514 0.086 0.029 0.000
##
## Node number 1812: 405 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5012346 P(node) =0.001473783
## class counts: 202 140 42 18 3
## probabilities: 0.499 0.346 0.104 0.044 0.007
## left son=3624 (329 obs) right son=3625 (76 obs)
## Primary splits:
## age < 83.5 to the left, improve=1.8474760, (0 missing)
## reimbursement2008 < 14045 to the left, improve=1.4437850, (0 missing)
## kidney < 0.5 to the right, improve=1.0197570, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4573240, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4260458, (0 missing)
##
## Node number 1813: 452 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.579646 P(node) =0.001644815
## class counts: 190 173 58 30 1
## probabilities: 0.420 0.383 0.128 0.066 0.002
## left son=3626 (362 obs) right son=3627 (90 obs)
## Primary splits:
## reimbursement2008 < 3875 to the right, improve=2.645100, (0 missing)
## age < 84.5 to the left, improve=2.429876, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.612268, (0 missing)
## alzheimers < 0.5 to the right, improve=1.063100, (0 missing)
## kidney < 0.5 to the left, improve=0.663279, (0 missing)
## Surrogate splits:
## age < 32 to the right, agree=0.803, adj=0.011, (0 split)
##
## Node number 1820: 177 observations
## predicted class=B1 expected loss=0.559322 P(node) =0.0006440978
## class counts: 78 62 26 11 0
## probabilities: 0.441 0.350 0.147 0.062 0.000
##
## Node number 1821: 519 observations
## predicted class=B2 expected loss=0.6088632 P(node) =0.001888626
## class counts: 154 203 86 64 12
## probabilities: 0.297 0.391 0.166 0.123 0.023
##
## Node number 1826: 25 observations
## predicted class=B1 expected loss=0.32 P(node) =9.097426e-05
## class counts: 17 7 1 0 0
## probabilities: 0.680 0.280 0.040 0.000 0.000
##
## Node number 1827: 357 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5826331 P(node) =0.001299112
## class counts: 140 149 48 19 1
## probabilities: 0.392 0.417 0.134 0.053 0.003
## left son=3654 (91 obs) right son=3655 (266 obs)
## Primary splits:
## age < 80.5 to the right, improve=2.1730570, (0 missing)
## reimbursement2008 < 4405 to the right, improve=2.0106540, (0 missing)
## depression < 0.5 to the left, improve=0.7793758, (0 missing)
## kidney < 0.5 to the left, improve=0.7738464, (0 missing)
## alzheimers < 0.5 to the right, improve=0.6298514, (0 missing)
##
## Node number 1846: 39 observations
## predicted class=B1 expected loss=0.4871795 P(node) =0.0001419198
## class counts: 20 12 4 3 0
## probabilities: 0.513 0.308 0.103 0.077 0.000
##
## Node number 1847: 254 observations
## predicted class=B2 expected loss=0.5787402 P(node) =0.0009242985
## class counts: 76 107 44 27 0
## probabilities: 0.299 0.421 0.173 0.106 0.000
##
## Node number 1848: 132 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5909091 P(node) =0.0004803441
## class counts: 54 42 26 9 1
## probabilities: 0.409 0.318 0.197 0.068 0.008
## left son=3696 (105 obs) right son=3697 (27 obs)
## Primary splits:
## age < 84.5 to the left, improve=4.2569990, (0 missing)
## reimbursement2008 < 13440 to the left, improve=1.5425260, (0 missing)
## ihd < 0.5 to the right, improve=1.3693110, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4363743, (0 missing)
## kidney < 0.5 to the left, improve=0.4353832, (0 missing)
##
## Node number 1849: 88 observations
## predicted class=B3 expected loss=0.6477273 P(node) =0.0003202294
## class counts: 23 24 31 7 3
## probabilities: 0.261 0.273 0.352 0.080 0.034
##
## Node number 1898: 54 observations
## predicted class=B3 expected loss=0.5555556 P(node) =0.0001965044
## class counts: 8 14 24 7 1
## probabilities: 0.148 0.259 0.444 0.130 0.019
##
## Node number 1899: 57 observations
## predicted class=B2 expected loss=0.5614035 P(node) =0.0002074213
## class counts: 5 25 15 10 2
## probabilities: 0.088 0.439 0.263 0.175 0.035
##
## Node number 1924: 1127 observations
## predicted class=B1 expected loss=0.4960071 P(node) =0.00410112
## class counts: 568 377 131 47 4
## probabilities: 0.504 0.335 0.116 0.042 0.004
##
## Node number 1925: 509 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5658153 P(node) =0.001852236
## class counts: 221 185 67 33 3
## probabilities: 0.434 0.363 0.132 0.065 0.006
## left son=3850 (137 obs) right son=3851 (372 obs)
## Primary splits:
## reimbursement2008 < 3775 to the left, improve=1.6880360, (0 missing)
## depression < 0.5 to the left, improve=1.6361880, (0 missing)
## age < 96.5 to the left, improve=1.5026800, (0 missing)
## copd < 0.5 to the left, improve=1.2566690, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7418596, (0 missing)
##
## Node number 1926: 339 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5575221 P(node) =0.001233611
## class counts: 150 127 45 16 1
## probabilities: 0.442 0.375 0.133 0.047 0.003
## left son=3852 (211 obs) right son=3853 (128 obs)
## Primary splits:
## reimbursement2008 < 4905 to the left, improve=1.6963240, (0 missing)
## age < 45 to the right, improve=1.4829560, (0 missing)
## heart.failure < 0.5 to the left, improve=1.2573130, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6055009, (0 missing)
## copd < 0.5 to the left, improve=0.2708942, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.687, adj=0.172, (0 split)
## copd < 0.5 to the left, agree=0.664, adj=0.109, (0 split)
## stroke < 0.5 to the left, agree=0.652, adj=0.078, (0 split)
##
## Node number 1927: 239 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.5439331 P(node) =0.0008697139
## class counts: 70 109 47 11 2
## probabilities: 0.293 0.456 0.197 0.046 0.008
## left son=3854 (181 obs) right son=3855 (58 obs)
## Primary splits:
## copd < 0.5 to the left, improve=6.3468870, (0 missing)
## reimbursement2008 < 5790 to the right, improve=1.7891020, (0 missing)
## age < 60.5 to the left, improve=1.2691270, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8740764, (0 missing)
## stroke < 0.5 to the right, improve=0.6684821, (0 missing)
## Surrogate splits:
## age < 35 to the right, agree=0.762, adj=0.017, (0 split)
##
## Node number 1928: 798 observations
## predicted class=B1 expected loss=0.5075188 P(node) =0.002903898
## class counts: 393 256 95 51 3
## probabilities: 0.492 0.321 0.119 0.064 0.004
##
## Node number 1929: 565 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.6247788 P(node) =0.002056018
## class counts: 212 179 124 41 9
## probabilities: 0.375 0.317 0.219 0.073 0.016
## left son=3858 (116 obs) right son=3859 (449 obs)
## Primary splits:
## stroke < 0.5 to the right, improve=4.0381830, (0 missing)
## reimbursement2008 < 31655 to the left, improve=2.7523450, (0 missing)
## age < 42.5 to the right, improve=1.7655450, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.4692280, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5615109, (0 missing)
## Surrogate splits:
## reimbursement2008 < 61780 to the right, agree=0.8, adj=0.026, (0 split)
##
## Node number 1930: 1953 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.578085 P(node) =0.007106909
## class counts: 824 782 251 88 8
## probabilities: 0.422 0.400 0.129 0.045 0.004
## left son=3860 (343 obs) right son=3861 (1610 obs)
## Primary splits:
## reimbursement2008 < 3415 to the left, improve=3.4037160, (0 missing)
## age < 42.5 to the left, improve=3.2783080, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6509623, (0 missing)
## copd < 0.5 to the left, improve=0.5598170, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2946889, (0 missing)
##
## Node number 1931: 2247 observations, complexity param=0.0001605101
## predicted class=B2 expected loss=0.5959057 P(node) =0.008176767
## class counts: 795 908 383 148 13
## probabilities: 0.354 0.404 0.170 0.066 0.006
## left son=3862 (866 obs) right son=3863 (1381 obs)
## Primary splits:
## reimbursement2008 < 5335 to the right, improve=3.344298, (0 missing)
## copd < 0.5 to the left, improve=2.798571, (0 missing)
## age < 68.5 to the left, improve=2.255236, (0 missing)
## alzheimers < 0.5 to the left, improve=1.653597, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.448247, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.678, adj=0.165, (0 split)
## age < 34.5 to the left, agree=0.616, adj=0.005, (0 split)
##
## Node number 1932: 1987 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5546049 P(node) =0.007230634
## class counts: 662 885 320 111 9
## probabilities: 0.333 0.445 0.161 0.056 0.005
## left son=3864 (1964 obs) right son=3865 (23 obs)
## Primary splits:
## age < 98.5 to the left, improve=3.328502, (0 missing)
## heart.failure < 0.5 to the left, improve=3.141909, (0 missing)
## reimbursement2008 < 3085 to the left, improve=3.126917, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.906536, (0 missing)
## alzheimers < 0.5 to the left, improve=1.332632, (0 missing)
##
## Node number 1933: 941 observations
## predicted class=B2 expected loss=0.5494155 P(node) =0.003424271
## class counts: 242 424 186 79 10
## probabilities: 0.257 0.451 0.198 0.084 0.011
##
## Node number 1934: 36 observations
## predicted class=B1 expected loss=0.4444444 P(node) =0.0001310029
## class counts: 20 8 8 0 0
## probabilities: 0.556 0.222 0.222 0.000 0.000
##
## Node number 1935: 1428 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6239496 P(node) =0.00519645
## class counts: 455 537 282 146 8
## probabilities: 0.319 0.376 0.197 0.102 0.006
## left son=3870 (837 obs) right son=3871 (591 obs)
## Primary splits:
## age < 78.5 to the left, improve=2.474561, (0 missing)
## heart.failure < 0.5 to the left, improve=2.118405, (0 missing)
## stroke < 0.5 to the left, improve=1.930317, (0 missing)
## copd < 0.5 to the left, improve=1.447977, (0 missing)
## reimbursement2008 < 8670 to the right, improve=1.324274, (0 missing)
##
## Node number 1984: 101 observations
## predicted class=B2 expected loss=0.5247525 P(node) =0.000367536
## class counts: 33 48 15 4 1
## probabilities: 0.327 0.475 0.149 0.040 0.010
##
## Node number 1985: 471 observations, complexity param=5.811572e-05
## predicted class=B1 expected loss=0.5732484 P(node) =0.001713955
## class counts: 201 135 77 53 5
## probabilities: 0.427 0.287 0.163 0.113 0.011
## left son=3970 (346 obs) right son=3971 (125 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.506365, (0 missing)
## reimbursement2008 < 11515 to the left, improve=2.004779, (0 missing)
## heart.failure < 0.5 to the left, improve=1.922393, (0 missing)
## age < 72.5 to the right, improve=1.840715, (0 missing)
## bucket2008 < 2.5 to the left, improve=1.312872, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3600 to the right, agree=0.737, adj=0.008, (0 split)
##
## Node number 1986: 9 observations
## predicted class=B1 expected loss=0.2222222 P(node) =3.275073e-05
## class counts: 7 2 0 0 0
## probabilities: 0.778 0.222 0.000 0.000 0.000
##
## Node number 1987: 383 observations
## predicted class=B2 expected loss=0.5744125 P(node) =0.001393726
## class counts: 115 163 75 25 5
## probabilities: 0.300 0.426 0.196 0.065 0.013
##
## Node number 1990: 2424 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6575908 P(node) =0.008820864
## class counts: 663 830 547 335 49
## probabilities: 0.274 0.342 0.226 0.138 0.020
## left son=3980 (1234 obs) right son=3981 (1190 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.612677, (0 missing)
## age < 67.5 to the right, improve=3.329297, (0 missing)
## copd < 0.5 to the left, improve=3.109296, (0 missing)
## heart.failure < 0.5 to the left, improve=2.737555, (0 missing)
## reimbursement2008 < 9205 to the right, improve=2.610291, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.552, adj=0.087, (0 split)
## copd < 0.5 to the left, agree=0.540, adj=0.064, (0 split)
## age < 53.5 to the right, agree=0.526, adj=0.035, (0 split)
## reimbursement2008 < 12525 to the left, agree=0.522, adj=0.026, (0 split)
## heart.failure < 0.5 to the left, agree=0.520, adj=0.023, (0 split)
##
## Node number 1991: 1226 observations
## predicted class=B2 expected loss=0.5831974 P(node) =0.004461378
## class counts: 254 511 265 175 21
## probabilities: 0.207 0.417 0.216 0.143 0.017
##
## Node number 2018: 293 observations
## predicted class=B2 expected loss=0.4914676 P(node) =0.001066218
## class counts: 20 149 81 39 4
## probabilities: 0.068 0.509 0.276 0.133 0.014
##
## Node number 2019: 25 observations
## predicted class=B3 expected loss=0.4 P(node) =9.097426e-05
## class counts: 3 1 15 6 0
## probabilities: 0.120 0.040 0.600 0.240 0.000
##
## Node number 2030: 237 observations, complexity param=6.088314e-05
## predicted class=B4 expected loss=0.6329114 P(node) =0.000862436
## class counts: 13 76 49 87 12
## probabilities: 0.055 0.321 0.207 0.367 0.051
## left son=4060 (62 obs) right son=4061 (175 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=2.618202, (0 missing)
## reimbursement2008 < 90420 to the right, improve=2.488954, (0 missing)
## heart.failure < 0.5 to the left, improve=2.039633, (0 missing)
## age < 61.5 to the left, improve=1.881916, (0 missing)
## alzheimers < 0.5 to the left, improve=1.753135, (0 missing)
##
## Node number 2031: 14 observations
## predicted class=B4 expected loss=0.2142857 P(node) =5.094559e-05
## class counts: 0 1 1 11 1
## probabilities: 0.000 0.071 0.071 0.786 0.071
##
## Node number 2032: 72 observations
## predicted class=B1 expected loss=0.5694444 P(node) =0.0002620059
## class counts: 31 18 11 8 4
## probabilities: 0.431 0.250 0.153 0.111 0.056
##
## Node number 2033: 1245 observations
## predicted class=B2 expected loss=0.6714859 P(node) =0.004530518
## class counts: 238 409 223 305 70
## probabilities: 0.191 0.329 0.179 0.245 0.056
##
## Node number 2034: 191 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6753927 P(node) =0.0006950434
## class counts: 29 62 44 42 14
## probabilities: 0.152 0.325 0.230 0.220 0.073
## left son=4068 (172 obs) right son=4069 (19 obs)
## Primary splits:
## age < 64.5 to the right, improve=2.5420870, (0 missing)
## reimbursement2008 < 71460 to the left, improve=2.3514440, (0 missing)
## alzheimers < 0.5 to the left, improve=1.2597430, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9221356, (0 missing)
## ihd < 0.5 to the right, improve=0.7384918, (0 missing)
##
## Node number 2035: 981 observations, complexity param=9.962695e-05
## predicted class=B4 expected loss=0.6788991 P(node) =0.00356983
## class counts: 243 203 158 315 62
## probabilities: 0.248 0.207 0.161 0.321 0.063
## left son=4070 (468 obs) right son=4071 (513 obs)
## Primary splits:
## reimbursement2008 < 23175 to the left, improve=5.196818, (0 missing)
## alzheimers < 0.5 to the right, improve=3.174409, (0 missing)
## ihd < 0.5 to the left, improve=2.640760, (0 missing)
## heart.failure < 0.5 to the left, improve=1.689264, (0 missing)
## age < 97.5 to the right, improve=1.688586, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.777, adj=0.532, (0 split)
## heart.failure < 0.5 to the left, agree=0.534, adj=0.024, (0 split)
## age < 53.5 to the left, agree=0.531, adj=0.017, (0 split)
## stroke < 0.5 to the right, agree=0.525, adj=0.004, (0 split)
##
## Node number 3304: 265 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.490566 P(node) =0.0009643272
## class counts: 135 82 34 13 1
## probabilities: 0.509 0.309 0.128 0.049 0.004
## left son=6608 (251 obs) right son=6609 (14 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=3.807509, (0 missing)
## ihd < 0.5 to the left, improve=2.996787, (0 missing)
## cancer < 0.5 to the left, improve=2.863288, (0 missing)
## reimbursement2008 < 1815 to the left, improve=1.417998, (0 missing)
## depression < 0.5 to the left, improve=1.227469, (0 missing)
##
## Node number 3305: 290 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5931034 P(node) =0.001055301
## class counts: 118 111 42 16 3
## probabilities: 0.407 0.383 0.145 0.055 0.010
## left son=6610 (213 obs) right son=6611 (77 obs)
## Primary splits:
## age < 81.5 to the left, improve=1.9355560, (0 missing)
## reimbursement2008 < 2015 to the right, improve=1.1719950, (0 missing)
## ihd < 0.5 to the right, improve=0.8443893, (0 missing)
## stroke < 0.5 to the right, improve=0.5640543, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4757090, (0 missing)
##
## Node number 3470: 840 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5452381 P(node) =0.003056735
## class counts: 382 309 103 41 5
## probabilities: 0.455 0.368 0.123 0.049 0.006
## left son=6940 (71 obs) right son=6941 (769 obs)
## Primary splits:
## age < 54.5 to the left, improve=2.5793890, (0 missing)
## kidney < 0.5 to the left, improve=2.0319410, (0 missing)
## depression < 0.5 to the left, improve=1.3091120, (0 missing)
## reimbursement2008 < 2945 to the right, improve=1.1530320, (0 missing)
## copd < 0.5 to the left, improve=0.9762638, (0 missing)
##
## Node number 3471: 217 observations
## predicted class=B2 expected loss=0.5345622 P(node) =0.0007896566
## class counts: 76 101 34 6 0
## probabilities: 0.350 0.465 0.157 0.028 0.000
##
## Node number 3500: 1099 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.5814377 P(node) =0.003999229
## class counts: 460 388 168 76 7
## probabilities: 0.419 0.353 0.153 0.069 0.006
## left son=7000 (1074 obs) right son=7001 (25 obs)
## Primary splits:
## age < 95.5 to the left, improve=2.515661, (0 missing)
## copd < 0.5 to the left, improve=2.359857, (0 missing)
## cancer < 0.5 to the left, improve=1.641148, (0 missing)
## reimbursement2008 < 2575 to the right, improve=1.347245, (0 missing)
## stroke < 0.5 to the left, improve=1.145174, (0 missing)
##
## Node number 3501: 653 observations, complexity param=0.0001129105
## predicted class=B2 expected loss=0.6309342 P(node) =0.002376248
## class counts: 226 241 126 55 5
## probabilities: 0.346 0.369 0.193 0.084 0.008
## left son=7002 (303 obs) right son=7003 (350 obs)
## Primary splits:
## reimbursement2008 < 2655 to the left, improve=2.636734, (0 missing)
## cancer < 0.5 to the left, improve=1.461370, (0 missing)
## age < 55.5 to the right, improve=1.350106, (0 missing)
## alzheimers < 0.5 to the left, improve=1.189997, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.091246, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.548, adj=0.026, (0 split)
## copd < 0.5 to the right, agree=0.542, adj=0.013, (0 split)
## age < 47.5 to the left, agree=0.539, adj=0.007, (0 split)
##
## Node number 3508: 382 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.6230366 P(node) =0.001390087
## class counts: 142 144 74 18 4
## probabilities: 0.372 0.377 0.194 0.047 0.010
## left son=7016 (229 obs) right son=7017 (153 obs)
## Primary splits:
## depression < 0.5 to the left, improve=0.9679873, (0 missing)
## reimbursement2008 < 2275 to the left, improve=0.8225283, (0 missing)
## age < 75.5 to the right, improve=0.7274055, (0 missing)
## cancer < 0.5 to the right, improve=0.6895810, (0 missing)
## alzheimers < 0.5 to the right, improve=0.4333447, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.620, adj=0.052, (0 split)
## age < 50.5 to the right, agree=0.613, adj=0.033, (0 split)
##
## Node number 3509: 21 observations
## predicted class=B1 expected loss=0.5714286 P(node) =7.641838e-05
## class counts: 9 3 5 2 2
## probabilities: 0.429 0.143 0.238 0.095 0.095
##
## Node number 3532: 322 observations
## predicted class=B1 expected loss=0.5465839 P(node) =0.001171748
## class counts: 146 119 43 14 0
## probabilities: 0.453 0.370 0.134 0.043 0.000
##
## Node number 3533: 14 observations
## predicted class=B2 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 4 9 1 0 0
## probabilities: 0.286 0.643 0.071 0.000 0.000
##
## Node number 3544: 1008 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.5853175 P(node) =0.003668082
## class counts: 418 400 133 53 4
## probabilities: 0.415 0.397 0.132 0.053 0.004
## left son=7088 (275 obs) right son=7089 (733 obs)
## Primary splits:
## reimbursement2008 < 2535 to the right, improve=0.9732083, (0 missing)
## age < 39 to the left, improve=0.9699606, (0 missing)
## copd < 0.5 to the left, improve=0.8468269, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4615681, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4416739, (0 missing)
## Surrogate splits:
## age < 36.5 to the left, agree=0.728, adj=0.004, (0 split)
##
## Node number 3545: 49 observations
## predicted class=B2 expected loss=0.5918367 P(node) =0.0001783096
## class counts: 16 20 12 1 0
## probabilities: 0.327 0.408 0.245 0.020 0.000
##
## Node number 3604: 286 observations
## predicted class=B1 expected loss=0.4685315 P(node) =0.001040746
## class counts: 152 94 35 5 0
## probabilities: 0.531 0.329 0.122 0.017 0.000
##
## Node number 3605: 398 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5477387 P(node) =0.00144831
## class counts: 180 137 60 21 0
## probabilities: 0.452 0.344 0.151 0.053 0.000
## left son=7210 (340 obs) right son=7211 (58 obs)
## Primary splits:
## reimbursement2008 < 4700 to the right, improve=5.7797840, (0 missing)
## alzheimers < 0.5 to the left, improve=1.1372610, (0 missing)
## age < 34.5 to the left, improve=0.9964329, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.5848011, (0 missing)
## kidney < 0.5 to the right, improve=0.4151452, (0 missing)
##
## Node number 3624: 329 observations
## predicted class=B1 expected loss=0.4832827 P(node) =0.001197221
## class counts: 170 105 35 16 3
## probabilities: 0.517 0.319 0.106 0.049 0.009
##
## Node number 3625: 76 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5394737 P(node) =0.0002765618
## class counts: 32 35 7 2 0
## probabilities: 0.421 0.461 0.092 0.026 0.000
## left son=7250 (21 obs) right son=7251 (55 obs)
## Primary splits:
## reimbursement2008 < 6785 to the right, improve=3.2066300, (0 missing)
## bucket2008 < 2.5 to the right, improve=3.1159910, (0 missing)
## alzheimers < 0.5 to the right, improve=1.9967220, (0 missing)
## age < 85.5 to the right, improve=1.1176690, (0 missing)
## kidney < 0.5 to the right, improve=0.9258269, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.921, adj=0.714, (0 split)
## kidney < 0.5 to the right, agree=0.789, adj=0.238, (0 split)
##
## Node number 3626: 362 observations
## predicted class=B1 expected loss=0.5552486 P(node) =0.001317307
## class counts: 161 128 47 25 1
## probabilities: 0.445 0.354 0.130 0.069 0.003
##
## Node number 3627: 90 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5 P(node) =0.0003275073
## class counts: 29 45 11 5 0
## probabilities: 0.322 0.500 0.122 0.056 0.000
## left son=7254 (21 obs) right son=7255 (69 obs)
## Primary splits:
## age < 69.5 to the left, improve=3.1544510, (0 missing)
## alzheimers < 0.5 to the right, improve=3.1535260, (0 missing)
## kidney < 0.5 to the left, improve=1.7000000, (0 missing)
## reimbursement2008 < 3185 to the left, improve=1.5133190, (0 missing)
## copd < 0.5 to the right, improve=0.3083333, (0 missing)
##
## Node number 3654: 91 observations
## predicted class=B1 expected loss=0.5054945 P(node) =0.0003311463
## class counts: 45 31 9 6 0
## probabilities: 0.495 0.341 0.099 0.066 0.000
##
## Node number 3655: 266 observations
## predicted class=B2 expected loss=0.556391 P(node) =0.0009679661
## class counts: 95 118 39 13 1
## probabilities: 0.357 0.444 0.147 0.049 0.004
##
## Node number 3696: 105 observations
## predicted class=B1 expected loss=0.5238095 P(node) =0.0003820919
## class counts: 50 35 16 3 1
## probabilities: 0.476 0.333 0.152 0.029 0.010
##
## Node number 3697: 27 observations
## predicted class=B3 expected loss=0.6296296 P(node) =9.82522e-05
## class counts: 4 7 10 6 0
## probabilities: 0.148 0.259 0.370 0.222 0.000
##
## Node number 3850: 137 observations
## predicted class=B1 expected loss=0.4963504 P(node) =0.000498539
## class counts: 69 41 17 9 1
## probabilities: 0.504 0.299 0.124 0.066 0.007
##
## Node number 3851: 372 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5913978 P(node) =0.001353697
## class counts: 152 144 50 24 2
## probabilities: 0.409 0.387 0.134 0.065 0.005
## left son=7702 (330 obs) right son=7703 (42 obs)
## Primary splits:
## reimbursement2008 < 4055 to the right, improve=3.5107950, (0 missing)
## age < 96 to the left, improve=2.0766610, (0 missing)
## depression < 0.5 to the left, improve=1.3295540, (0 missing)
## copd < 0.5 to the left, improve=1.2612770, (0 missing)
## heart.failure < 0.5 to the left, improve=0.1950857, (0 missing)
##
## Node number 3852: 211 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.563981 P(node) =0.0007678228
## class counts: 92 89 23 7 0
## probabilities: 0.436 0.422 0.109 0.033 0.000
## left son=7704 (142 obs) right son=7705 (69 obs)
## Primary splits:
## reimbursement2008 < 4075 to the left, improve=3.4884480, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9268848, (0 missing)
## heart.failure < 0.5 to the left, improve=0.8762896, (0 missing)
## age < 95 to the right, improve=0.6396384, (0 missing)
## stroke < 0.5 to the left, improve=0.3979516, (0 missing)
## Surrogate splits:
## age < 96.5 to the left, agree=0.687, adj=0.043, (0 split)
##
## Node number 3853: 128 observations
## predicted class=B1 expected loss=0.546875 P(node) =0.0004657882
## class counts: 58 38 22 9 1
## probabilities: 0.453 0.297 0.172 0.070 0.008
##
## Node number 3854: 181 observations
## predicted class=B2 expected loss=0.4861878 P(node) =0.0006586537
## class counts: 56 93 23 7 2
## probabilities: 0.309 0.514 0.127 0.039 0.011
##
## Node number 3855: 58 observations
## predicted class=B3 expected loss=0.5862069 P(node) =0.0002110603
## class counts: 14 16 24 4 0
## probabilities: 0.241 0.276 0.414 0.069 0.000
##
## Node number 3858: 116 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.5517241 P(node) =0.0004221206
## class counts: 41 52 14 7 2
## probabilities: 0.353 0.448 0.121 0.060 0.017
## left son=7716 (63 obs) right son=7717 (53 obs)
## Primary splits:
## age < 74.5 to the right, improve=2.8010500, (0 missing)
## reimbursement2008 < 17265 to the left, improve=2.4722090, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.9776500, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5892240, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5845524, (0 missing)
## Surrogate splits:
## reimbursement2008 < 15590 to the right, agree=0.603, adj=0.132, (0 split)
## heart.failure < 0.5 to the right, agree=0.578, adj=0.075, (0 split)
## alzheimers < 0.5 to the right, agree=0.560, adj=0.038, (0 split)
## osteoporosis < 0.5 to the left, agree=0.552, adj=0.019, (0 split)
## bucket2008 < 3.5 to the right, agree=0.552, adj=0.019, (0 split)
##
## Node number 3859: 449 observations
## predicted class=B1 expected loss=0.6191537 P(node) =0.001633898
## class counts: 171 127 110 34 7
## probabilities: 0.381 0.283 0.245 0.076 0.016
##
## Node number 3860: 343 observations
## predicted class=B1 expected loss=0.5014577 P(node) =0.001248167
## class counts: 171 126 33 11 2
## probabilities: 0.499 0.367 0.096 0.032 0.006
##
## Node number 3861: 1610 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5925466 P(node) =0.005858742
## class counts: 653 656 218 77 6
## probabilities: 0.406 0.407 0.135 0.048 0.004
## left son=7722 (43 obs) right son=7723 (1567 obs)
## Primary splits:
## age < 42.5 to the left, improve=2.9787580, (0 missing)
## reimbursement2008 < 3475 to the right, improve=1.6291410, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4315089, (0 missing)
## copd < 0.5 to the left, improve=0.2703192, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2492479, (0 missing)
##
## Node number 3862: 866 observations, complexity param=0.0001605101
## predicted class=B1 expected loss=0.6120092 P(node) =0.003151348
## class counts: 336 322 139 64 5
## probabilities: 0.388 0.372 0.161 0.074 0.006
## left son=7724 (129 obs) right son=7725 (737 obs)
## Primary splits:
## reimbursement2008 < 8115 to the right, improve=2.0175360, (0 missing)
## age < 89.5 to the left, improve=1.8274460, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.6689370, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.3462440, (0 missing)
## stroke < 0.5 to the left, improve=0.5970317, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.984, adj=0.891, (0 split)
##
## Node number 3863: 1381 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.5756698 P(node) =0.005025418
## class counts: 459 586 244 84 8
## probabilities: 0.332 0.424 0.177 0.061 0.006
## left son=7726 (997 obs) right son=7727 (384 obs)
## Primary splits:
## copd < 0.5 to the left, improve=3.5627620, (0 missing)
## age < 37.5 to the right, improve=2.2016010, (0 missing)
## reimbursement2008 < 5195 to the left, improve=1.9417980, (0 missing)
## alzheimers < 0.5 to the left, improve=1.9283600, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3841814, (0 missing)
## Surrogate splits:
## age < 34 to the right, agree=0.723, adj=0.005, (0 split)
## reimbursement2008 < 5325 to the left, agree=0.723, adj=0.005, (0 split)
##
## Node number 3864: 1964 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.552444 P(node) =0.007146938
## class counts: 656 879 309 111 9
## probabilities: 0.334 0.448 0.157 0.057 0.005
## left son=7728 (22 obs) right son=7729 (1942 obs)
## Primary splits:
## reimbursement2008 < 3085 to the left, improve=3.418849, (0 missing)
## heart.failure < 0.5 to the left, improve=3.308540, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.919418, (0 missing)
## age < 66.5 to the right, improve=1.961336, (0 missing)
## alzheimers < 0.5 to the left, improve=1.448295, (0 missing)
##
## Node number 3865: 23 observations
## predicted class=B3 expected loss=0.5217391 P(node) =8.369632e-05
## class counts: 6 6 11 0 0
## probabilities: 0.261 0.261 0.478 0.000 0.000
##
## Node number 3870: 837 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6356033 P(node) =0.003045818
## class counts: 292 305 155 82 3
## probabilities: 0.349 0.364 0.185 0.098 0.004
## left son=7740 (639 obs) right son=7741 (198 obs)
## Primary splits:
## reimbursement2008 < 21320 to the left, improve=1.8992410, (0 missing)
## alzheimers < 0.5 to the left, improve=1.6748470, (0 missing)
## age < 49.5 to the left, improve=1.4981360, (0 missing)
## bucket2008 < 3.5 to the left, improve=1.4460210, (0 missing)
## stroke < 0.5 to the right, improve=0.7571134, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.955, adj=0.808, (0 split)
##
## Node number 3871: 591 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.607445 P(node) =0.002150632
## class counts: 163 232 127 64 5
## probabilities: 0.276 0.393 0.215 0.108 0.008
## left son=7742 (122 obs) right son=7743 (469 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.4607800, (0 missing)
## reimbursement2008 < 8775 to the left, improve=1.9199300, (0 missing)
## stroke < 0.5 to the left, improve=1.2822570, (0 missing)
## copd < 0.5 to the left, improve=1.1325150, (0 missing)
## age < 80.5 to the right, improve=0.6677683, (0 missing)
##
## Node number 3970: 346 observations
## predicted class=B1 expected loss=0.5375723 P(node) =0.001259084
## class counts: 160 92 52 37 5
## probabilities: 0.462 0.266 0.150 0.107 0.014
##
## Node number 3971: 125 observations, complexity param=5.811572e-05
## predicted class=B2 expected loss=0.656 P(node) =0.0004548713
## class counts: 41 43 25 16 0
## probabilities: 0.328 0.344 0.200 0.128 0.000
## left son=7942 (106 obs) right son=7943 (19 obs)
## Primary splits:
## age < 62 to the right, improve=3.3415930, (0 missing)
## reimbursement2008 < 11475 to the left, improve=2.2730020, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7920000, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.7319750, (0 missing)
## stroke < 0.5 to the right, improve=0.5402967, (0 missing)
##
## Node number 3980: 1234 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6612642 P(node) =0.00449049
## class counts: 374 418 252 160 30
## probabilities: 0.303 0.339 0.204 0.130 0.024
## left son=7960 (349 obs) right son=7961 (885 obs)
## Primary splits:
## reimbursement2008 < 12135 to the right, improve=3.745241, (0 missing)
## age < 67.5 to the left, improve=3.421516, (0 missing)
## heart.failure < 0.5 to the left, improve=1.338981, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.254047, (0 missing)
## copd < 0.5 to the right, improve=1.093433, (0 missing)
##
## Node number 3981: 1190 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6537815 P(node) =0.004330375
## class counts: 289 412 295 175 19
## probabilities: 0.243 0.346 0.248 0.147 0.016
## left son=7962 (547 obs) right son=7963 (643 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.655589, (0 missing)
## age < 82.5 to the right, improve=1.853724, (0 missing)
## stroke < 0.5 to the right, improve=1.583996, (0 missing)
## reimbursement2008 < 6355 to the right, improve=1.325969, (0 missing)
## heart.failure < 0.5 to the left, improve=1.085887, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the left, agree=0.562, adj=0.048, (0 split)
## reimbursement2008 < 7315 to the left, agree=0.555, adj=0.031, (0 split)
## alzheimers < 0.5 to the left, agree=0.543, adj=0.005, (0 split)
## age < 28.5 to the left, agree=0.542, adj=0.004, (0 split)
##
## Node number 4060: 62 observations
## predicted class=B2 expected loss=0.5806452 P(node) =0.0002256162
## class counts: 3 26 17 15 1
## probabilities: 0.048 0.419 0.274 0.242 0.016
##
## Node number 4061: 175 observations
## predicted class=B4 expected loss=0.5885714 P(node) =0.0006368198
## class counts: 10 50 32 72 11
## probabilities: 0.057 0.286 0.183 0.411 0.063
##
## Node number 4068: 172 observations
## predicted class=B2 expected loss=0.6511628 P(node) =0.0006259029
## class counts: 27 60 35 39 11
## probabilities: 0.157 0.349 0.203 0.227 0.064
##
## Node number 4069: 19 observations
## predicted class=B3 expected loss=0.5263158 P(node) =6.914044e-05
## class counts: 2 2 9 3 3
## probabilities: 0.105 0.105 0.474 0.158 0.158
##
## Node number 4070: 468 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7200855 P(node) =0.001703038
## class counts: 131 112 77 122 26
## probabilities: 0.280 0.239 0.165 0.261 0.056
## left son=8140 (457 obs) right son=8141 (11 obs)
## Primary splits:
## age < 93.5 to the left, improve=1.955850, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.902288, (0 missing)
## reimbursement2008 < 19700 to the right, improve=1.873153, (0 missing)
## ihd < 0.5 to the right, improve=1.868954, (0 missing)
## alzheimers < 0.5 to the right, improve=1.716285, (0 missing)
##
## Node number 4071: 513 observations
## predicted class=B4 expected loss=0.6237817 P(node) =0.001866792
## class counts: 112 91 81 193 36
## probabilities: 0.218 0.177 0.158 0.376 0.070
##
## Node number 6608: 251 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.4701195 P(node) =0.0009133816
## class counts: 133 73 33 11 1
## probabilities: 0.530 0.291 0.131 0.044 0.004
## left son=13216 (235 obs) right son=13217 (16 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.6866090, (0 missing)
## ihd < 0.5 to the left, improve=2.4584220, (0 missing)
## reimbursement2008 < 1815 to the left, improve=1.2636760, (0 missing)
## age < 60.5 to the right, improve=1.0616730, (0 missing)
## depression < 0.5 to the left, improve=0.7871996, (0 missing)
##
## Node number 6609: 14 observations
## predicted class=B2 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 2 9 1 2 0
## probabilities: 0.143 0.643 0.071 0.143 0.000
##
## Node number 6610: 213 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5633803 P(node) =0.0007751007
## class counts: 93 74 30 15 1
## probabilities: 0.437 0.347 0.141 0.070 0.005
## left son=13220 (201 obs) right son=13221 (12 obs)
## Primary splits:
## age < 44.5 to the right, improve=1.7572700, (0 missing)
## reimbursement2008 < 2005 to the right, improve=1.0657280, (0 missing)
## cancer < 0.5 to the right, improve=0.3892571, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3640621, (0 missing)
## depression < 0.5 to the right, improve=0.3467310, (0 missing)
##
## Node number 6611: 77 observations
## predicted class=B2 expected loss=0.5194805 P(node) =0.0002802007
## class counts: 25 37 12 1 2
## probabilities: 0.325 0.481 0.156 0.013 0.026
##
## Node number 6940: 71 observations
## predicted class=B1 expected loss=0.4366197 P(node) =0.0002583669
## class counts: 40 16 11 3 1
## probabilities: 0.563 0.225 0.155 0.042 0.014
##
## Node number 6941: 769 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5552666 P(node) =0.002798368
## class counts: 342 293 92 38 4
## probabilities: 0.445 0.381 0.120 0.049 0.005
## left son=13882 (472 obs) right son=13883 (297 obs)
## Primary splits:
## age < 70.5 to the right, improve=2.5248320, (0 missing)
## depression < 0.5 to the left, improve=1.8557100, (0 missing)
## kidney < 0.5 to the left, improve=1.7236880, (0 missing)
## reimbursement2008 < 2665 to the right, improve=1.1252400, (0 missing)
## copd < 0.5 to the left, improve=0.9387137, (0 missing)
##
## Node number 7000: 1074 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5772812 P(node) =0.003908254
## class counts: 454 373 166 74 7
## probabilities: 0.423 0.347 0.155 0.069 0.007
## left son=14000 (808 obs) right son=14001 (266 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.7468870, (0 missing)
## cancer < 0.5 to the left, improve=1.8315690, (0 missing)
## age < 78.5 to the left, improve=1.6074350, (0 missing)
## reimbursement2008 < 2575 to the right, improve=1.2651380, (0 missing)
## stroke < 0.5 to the left, improve=0.9951466, (0 missing)
##
## Node number 7001: 25 observations
## predicted class=B2 expected loss=0.4 P(node) =9.097426e-05
## class counts: 6 15 2 2 0
## probabilities: 0.240 0.600 0.080 0.080 0.000
##
## Node number 7002: 303 observations
## predicted class=B1 expected loss=0.6039604 P(node) =0.001102608
## class counts: 120 99 62 21 1
## probabilities: 0.396 0.327 0.205 0.069 0.003
##
## Node number 7003: 350 observations
## predicted class=B2 expected loss=0.5942857 P(node) =0.00127364
## class counts: 106 142 64 34 4
## probabilities: 0.303 0.406 0.183 0.097 0.011
##
## Node number 7016: 229 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5938865 P(node) =0.0008333242
## class counts: 93 82 44 8 2
## probabilities: 0.406 0.358 0.192 0.035 0.009
## left son=14032 (15 obs) right son=14033 (214 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=1.9222650, (0 missing)
## reimbursement2008 < 2515 to the left, improve=1.4164780, (0 missing)
## age < 94 to the left, improve=1.2547820, (0 missing)
## alzheimers < 0.5 to the right, improve=0.6631197, (0 missing)
## copd < 0.5 to the right, improve=0.2469242, (0 missing)
##
## Node number 7017: 153 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.5947712 P(node) =0.0005567625
## class counts: 49 62 30 10 2
## probabilities: 0.320 0.405 0.196 0.065 0.013
## left son=14034 (14 obs) right son=14035 (139 obs)
## Primary splits:
## reimbursement2008 < 2545 to the right, improve=2.7113570, (0 missing)
## age < 45 to the left, improve=1.6972360, (0 missing)
## cancer < 0.5 to the left, improve=0.6348039, (0 missing)
## copd < 0.5 to the right, improve=0.3887797, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2839287, (0 missing)
##
## Node number 7088: 275 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.56 P(node) =0.001000717
## class counts: 108 121 34 11 1
## probabilities: 0.393 0.440 0.124 0.040 0.004
## left son=14176 (44 obs) right son=14177 (231 obs)
## Primary splits:
## age < 63.5 to the left, improve=2.2068400, (0 missing)
## reimbursement2008 < 2555 to the right, improve=1.7374730, (0 missing)
## alzheimers < 0.5 to the right, improve=1.5968660, (0 missing)
## copd < 0.5 to the left, improve=0.9046397, (0 missing)
## heart.failure < 0.5 to the right, improve=0.5279104, (0 missing)
##
## Node number 7089: 733 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5770805 P(node) =0.002667365
## class counts: 310 279 99 42 3
## probabilities: 0.423 0.381 0.135 0.057 0.004
## left son=14178 (10 obs) right son=14179 (723 obs)
## Primary splits:
## age < 97.5 to the right, improve=1.1550490, (0 missing)
## heart.failure < 0.5 to the left, improve=1.1107930, (0 missing)
## reimbursement2008 < 2495 to the right, improve=0.7495829, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7242328, (0 missing)
## depression < 0.5 to the left, improve=0.5684301, (0 missing)
##
## Node number 7210: 340 observations
## predicted class=B1 expected loss=0.5088235 P(node) =0.00123725
## class counts: 167 107 48 18 0
## probabilities: 0.491 0.315 0.141 0.053 0.000
##
## Node number 7211: 58 observations
## predicted class=B2 expected loss=0.4827586 P(node) =0.0002110603
## class counts: 13 30 12 3 0
## probabilities: 0.224 0.517 0.207 0.052 0.000
##
## Node number 7250: 21 observations
## predicted class=B1 expected loss=0.3333333 P(node) =7.641838e-05
## class counts: 14 5 2 0 0
## probabilities: 0.667 0.238 0.095 0.000 0.000
##
## Node number 7251: 55 observations
## predicted class=B2 expected loss=0.4545455 P(node) =0.0002001434
## class counts: 18 30 5 2 0
## probabilities: 0.327 0.545 0.091 0.036 0.000
##
## Node number 7254: 21 observations
## predicted class=B1 expected loss=0.4285714 P(node) =7.641838e-05
## class counts: 12 6 1 2 0
## probabilities: 0.571 0.286 0.048 0.095 0.000
##
## Node number 7255: 69 observations
## predicted class=B2 expected loss=0.4347826 P(node) =0.000251089
## class counts: 17 39 10 3 0
## probabilities: 0.246 0.565 0.145 0.043 0.000
##
## Node number 7702: 330 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.569697 P(node) =0.00120086
## class counts: 142 119 44 23 2
## probabilities: 0.430 0.361 0.133 0.070 0.006
## left son=15404 (309 obs) right son=15405 (21 obs)
## Primary splits:
## reimbursement2008 < 4185 to the right, improve=2.2681710, (0 missing)
## age < 96 to the left, improve=2.1333520, (0 missing)
## depression < 0.5 to the left, improve=0.7533962, (0 missing)
## copd < 0.5 to the left, improve=0.6700147, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.3769465, (0 missing)
##
## Node number 7703: 42 observations
## predicted class=B2 expected loss=0.4047619 P(node) =0.0001528368
## class counts: 10 25 6 1 0
## probabilities: 0.238 0.595 0.143 0.024 0.000
##
## Node number 7704: 142 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0005167338
## class counts: 71 51 15 5 0
## probabilities: 0.500 0.359 0.106 0.035 0.000
##
## Node number 7705: 69 observations
## predicted class=B2 expected loss=0.4492754 P(node) =0.000251089
## class counts: 21 38 8 2 0
## probabilities: 0.304 0.551 0.116 0.029 0.000
##
## Node number 7716: 63 observations
## predicted class=B1 expected loss=0.5714286 P(node) =0.0002292551
## class counts: 27 21 10 4 1
## probabilities: 0.429 0.333 0.159 0.063 0.016
##
## Node number 7717: 53 observations
## predicted class=B2 expected loss=0.4150943 P(node) =0.0001928654
## class counts: 14 31 4 3 1
## probabilities: 0.264 0.585 0.075 0.057 0.019
##
## Node number 7722: 43 observations
## predicted class=B1 expected loss=0.3953488 P(node) =0.0001564757
## class counts: 26 11 3 3 0
## probabilities: 0.605 0.256 0.070 0.070 0.000
##
## Node number 7723: 1567 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5883854 P(node) =0.005702267
## class counts: 627 645 215 74 6
## probabilities: 0.400 0.412 0.137 0.047 0.004
## left son=15446 (1527 obs) right son=15447 (40 obs)
## Primary splits:
## age < 50.5 to the right, improve=1.7032880, (0 missing)
## reimbursement2008 < 3475 to the right, improve=1.5459000, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4552758, (0 missing)
## stroke < 0.5 to the left, improve=0.2471234, (0 missing)
## copd < 0.5 to the left, improve=0.2014160, (0 missing)
##
## Node number 7724: 129 observations
## predicted class=B2 expected loss=0.5271318 P(node) =0.0004694272
## class counts: 46 61 16 6 0
## probabilities: 0.357 0.473 0.124 0.047 0.000
##
## Node number 7725: 737 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.6065129 P(node) =0.002681921
## class counts: 290 261 123 58 5
## probabilities: 0.393 0.354 0.167 0.079 0.007
## left son=15450 (703 obs) right son=15451 (34 obs)
## Primary splits:
## age < 94.5 to the left, improve=1.9050170, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.5567680, (0 missing)
## reimbursement2008 < 6575 to the right, improve=1.5078350, (0 missing)
## copd < 0.5 to the left, improve=0.5423379, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5213862, (0 missing)
##
## Node number 7726: 997 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.5927783 P(node) =0.003628054
## class counts: 357 406 171 57 6
## probabilities: 0.358 0.407 0.172 0.057 0.006
## left son=15452 (297 obs) right son=15453 (700 obs)
## Primary splits:
## age < 69.5 to the left, improve=2.4458440, (0 missing)
## alzheimers < 0.5 to the left, improve=2.2624190, (0 missing)
## reimbursement2008 < 4135 to the right, improve=1.8635870, (0 missing)
## stroke < 0.5 to the right, improve=0.3191114, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3114115, (0 missing)
## Surrogate splits:
## reimbursement2008 < 5325 to the right, agree=0.703, adj=0.003, (0 split)
##
## Node number 7727: 384 observations
## predicted class=B2 expected loss=0.53125 P(node) =0.001397365
## class counts: 102 180 73 27 2
## probabilities: 0.266 0.469 0.190 0.070 0.005
##
## Node number 7728: 22 observations
## predicted class=B1 expected loss=0.3636364 P(node) =8.005735e-05
## class counts: 14 5 1 2 0
## probabilities: 0.636 0.227 0.045 0.091 0.000
##
## Node number 7729: 1942 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5499485 P(node) =0.007066881
## class counts: 642 874 308 109 9
## probabilities: 0.331 0.450 0.159 0.056 0.005
## left son=15458 (889 obs) right son=15459 (1053 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.215424, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.693064, (0 missing)
## reimbursement2008 < 8025 to the left, improve=2.054172, (0 missing)
## age < 66.5 to the right, improve=1.953237, (0 missing)
## alzheimers < 0.5 to the left, improve=1.238773, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3815 to the left, agree=0.545, adj=0.007, (0 split)
## age < 31.5 to the left, agree=0.544, adj=0.003, (0 split)
##
## Node number 7740: 639 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6353678 P(node) =0.002325302
## class counts: 233 221 124 59 2
## probabilities: 0.365 0.346 0.194 0.092 0.003
## left son=15480 (83 obs) right son=15481 (556 obs)
## Primary splits:
## age < 49.5 to the left, improve=1.7453130, (0 missing)
## stroke < 0.5 to the left, improve=1.0790130, (0 missing)
## reimbursement2008 < 10445 to the right, improve=1.0644190, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9731198, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.6917546, (0 missing)
##
## Node number 7741: 198 observations
## predicted class=B2 expected loss=0.5757576 P(node) =0.0007205162
## class counts: 59 84 31 23 1
## probabilities: 0.298 0.424 0.157 0.116 0.005
##
## Node number 7742: 122 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5983607 P(node) =0.0004439544
## class counts: 49 39 22 12 0
## probabilities: 0.402 0.320 0.180 0.098 0.000
## left son=15484 (40 obs) right son=15485 (82 obs)
## Primary splits:
## reimbursement2008 < 11560 to the left, improve=2.7817470, (0 missing)
## age < 80.5 to the right, improve=1.9103730, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4632510, (0 missing)
## copd < 0.5 to the left, improve=1.0641520, (0 missing)
## stroke < 0.5 to the left, improve=0.9890302, (0 missing)
##
## Node number 7743: 469 observations
## predicted class=B2 expected loss=0.5884861 P(node) =0.001706677
## class counts: 114 193 105 52 5
## probabilities: 0.243 0.412 0.224 0.111 0.011
##
## Node number 7942: 106 observations, complexity param=5.811572e-05
## predicted class=B1 expected loss=0.6320755 P(node) =0.0003857309
## class counts: 39 39 16 12 0
## probabilities: 0.368 0.368 0.151 0.113 0.000
## left son=15884 (93 obs) right son=15885 (13 obs)
## Primary splits:
## age < 67.5 to the right, improve=2.3273090, (0 missing)
## reimbursement2008 < 11575 to the left, improve=1.8244140, (0 missing)
## stroke < 0.5 to the right, improve=0.5034792, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4237564, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3937905, (0 missing)
##
## Node number 7943: 19 observations
## predicted class=B3 expected loss=0.5263158 P(node) =6.914044e-05
## class counts: 2 4 9 4 0
## probabilities: 0.105 0.211 0.474 0.211 0.000
##
## Node number 7960: 349 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6475645 P(node) =0.001270001
## class counts: 123 103 55 57 11
## probabilities: 0.352 0.295 0.158 0.163 0.032
## left son=15920 (331 obs) right son=15921 (18 obs)
## Primary splits:
## age < 54 to the right, improve=2.0196730, (0 missing)
## alzheimers < 0.5 to the left, improve=1.9958000, (0 missing)
## reimbursement2008 < 15235 to the left, improve=1.7314800, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9381122, (0 missing)
## copd < 0.5 to the right, improve=0.4818854, (0 missing)
##
## Node number 7961: 885 observations
## predicted class=B2 expected loss=0.6440678 P(node) =0.003220489
## class counts: 251 315 197 103 19
## probabilities: 0.284 0.356 0.223 0.116 0.021
##
## Node number 7962: 547 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6709324 P(node) =0.001990517
## class counts: 154 180 136 65 12
## probabilities: 0.282 0.329 0.249 0.119 0.022
## left son=15924 (310 obs) right son=15925 (237 obs)
## Primary splits:
## reimbursement2008 < 9205 to the right, improve=2.7787810, (0 missing)
## age < 68.5 to the right, improve=2.5123800, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.9624740, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6055315, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5192530, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.835, adj=0.620, (0 split)
## age < 44.5 to the right, agree=0.581, adj=0.034, (0 split)
##
## Node number 7963: 643 observations
## predicted class=B2 expected loss=0.6391913 P(node) =0.002339858
## class counts: 135 232 159 110 7
## probabilities: 0.210 0.361 0.247 0.171 0.011
##
## Node number 8140: 457 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7133479 P(node) =0.00166301
## class counts: 131 107 73 120 26
## probabilities: 0.287 0.234 0.160 0.263 0.057
## left son=16280 (398 obs) right son=16281 (59 obs)
## Primary splits:
## age < 86.5 to the left, improve=2.218362, (0 missing)
## ihd < 0.5 to the right, improve=2.044330, (0 missing)
## reimbursement2008 < 19430 to the right, improve=1.853412, (0 missing)
## alzheimers < 0.5 to the right, improve=1.735004, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.352826, (0 missing)
##
## Node number 8141: 11 observations
## predicted class=B2 expected loss=0.5454545 P(node) =4.002868e-05
## class counts: 0 5 4 2 0
## probabilities: 0.000 0.455 0.364 0.182 0.000
##
## Node number 13216: 235 observations
## predicted class=B1 expected loss=0.4510638 P(node) =0.0008551581
## class counts: 129 64 30 11 1
## probabilities: 0.549 0.272 0.128 0.047 0.004
##
## Node number 13217: 16 observations
## predicted class=B2 expected loss=0.4375 P(node) =5.822353e-05
## class counts: 4 9 3 0 0
## probabilities: 0.250 0.562 0.188 0.000 0.000
##
## Node number 13220: 201 observations
## predicted class=B1 expected loss=0.5472637 P(node) =0.0007314331
## class counts: 91 67 29 14 0
## probabilities: 0.453 0.333 0.144 0.070 0.000
##
## Node number 13221: 12 observations
## predicted class=B2 expected loss=0.4166667 P(node) =4.366765e-05
## class counts: 2 7 1 1 1
## probabilities: 0.167 0.583 0.083 0.083 0.083
##
## Node number 13882: 472 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5275424 P(node) =0.001717594
## class counts: 223 163 57 25 4
## probabilities: 0.472 0.345 0.121 0.053 0.008
## left son=27764 (343 obs) right son=27765 (129 obs)
## Primary splits:
## age < 73.5 to the right, improve=4.630612, (0 missing)
## reimbursement2008 < 2805 to the right, improve=1.597068, (0 missing)
## depression < 0.5 to the left, improve=1.459900, (0 missing)
## kidney < 0.5 to the left, improve=1.335760, (0 missing)
## stroke < 0.5 to the left, improve=1.130037, (0 missing)
##
## Node number 13883: 297 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.5622896 P(node) =0.001080774
## class counts: 119 130 35 13 0
## probabilities: 0.401 0.438 0.118 0.044 0.000
## left son=27766 (218 obs) right son=27767 (79 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=1.3951400, (0 missing)
## reimbursement2008 < 2945 to the right, improve=1.0350230, (0 missing)
## depression < 0.5 to the left, improve=0.9259259, (0 missing)
## kidney < 0.5 to the left, improve=0.7583938, (0 missing)
## copd < 0.5 to the left, improve=0.3569379, (0 missing)
##
## Node number 14000: 808 observations
## predicted class=B1 expected loss=0.5569307 P(node) =0.002940288
## class counts: 358 273 111 61 5
## probabilities: 0.443 0.338 0.137 0.075 0.006
##
## Node number 14001: 266 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.6240602 P(node) =0.0009679661
## class counts: 96 100 55 13 2
## probabilities: 0.361 0.376 0.207 0.049 0.008
## left son=28002 (192 obs) right son=28003 (74 obs)
## Primary splits:
## reimbursement2008 < 2540 to the right, improve=2.9691060, (0 missing)
## age < 78.5 to the left, improve=2.6852920, (0 missing)
## cancer < 0.5 to the right, improve=2.3754980, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7018574, (0 missing)
## stroke < 0.5 to the left, improve=0.6157537, (0 missing)
## Surrogate splits:
## age < 50.5 to the right, agree=0.737, adj=0.054, (0 split)
##
## Node number 14032: 15 observations
## predicted class=B1 expected loss=0.3333333 P(node) =5.458456e-05
## class counts: 10 2 3 0 0
## probabilities: 0.667 0.133 0.200 0.000 0.000
##
## Node number 14033: 214 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6121495 P(node) =0.0007787397
## class counts: 83 80 41 8 2
## probabilities: 0.388 0.374 0.192 0.037 0.009
## left son=28066 (169 obs) right son=28067 (45 obs)
## Primary splits:
## reimbursement2008 < 2515 to the left, improve=1.6030020, (0 missing)
## age < 52.5 to the right, improve=0.9765448, (0 missing)
## alzheimers < 0.5 to the right, improve=0.7668533, (0 missing)
## copd < 0.5 to the right, improve=0.3681910, (0 missing)
## ihd < 0.5 to the right, improve=0.1207875, (0 missing)
##
## Node number 14034: 14 observations
## predicted class=B1 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 9 2 2 1 0
## probabilities: 0.643 0.143 0.143 0.071 0.000
##
## Node number 14035: 139 observations
## predicted class=B2 expected loss=0.5683453 P(node) =0.0005058169
## class counts: 40 60 28 9 2
## probabilities: 0.288 0.432 0.201 0.065 0.014
##
## Node number 14176: 44 observations
## predicted class=B2 expected loss=0.3863636 P(node) =0.0001601147
## class counts: 14 27 2 1 0
## probabilities: 0.318 0.614 0.045 0.023 0.000
##
## Node number 14177: 231 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.5930736 P(node) =0.0008406022
## class counts: 94 94 32 10 1
## probabilities: 0.407 0.407 0.139 0.043 0.004
## left son=28354 (169 obs) right son=28355 (62 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.4583990, (0 missing)
## reimbursement2008 < 2555 to the right, improve=1.0376560, (0 missing)
## age < 84.5 to the left, improve=1.0243680, (0 missing)
## copd < 0.5 to the left, improve=0.7240171, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4410819, (0 missing)
## Surrogate splits:
## age < 87.5 to the left, agree=0.745, adj=0.048, (0 split)
##
## Node number 14178: 10 observations
## predicted class=B1 expected loss=0.3 P(node) =3.63897e-05
## class counts: 7 2 1 0 0
## probabilities: 0.700 0.200 0.100 0.000 0.000
##
## Node number 14179: 723 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5809129 P(node) =0.002630976
## class counts: 303 277 98 42 3
## probabilities: 0.419 0.383 0.136 0.058 0.004
## left son=28358 (689 obs) right son=28359 (34 obs)
## Primary splits:
## age < 90.5 to the left, improve=1.6650270, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5078050, (0 missing)
## reimbursement2008 < 2495 to the right, improve=0.8133392, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6699213, (0 missing)
## depression < 0.5 to the left, improve=0.5296598, (0 missing)
##
## Node number 15404: 309 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5728155 P(node) =0.001124442
## class counts: 132 117 38 20 2
## probabilities: 0.427 0.379 0.123 0.065 0.006
## left son=30808 (253 obs) right son=30809 (56 obs)
## Primary splits:
## reimbursement2008 < 4635 to the right, improve=2.0908250, (0 missing)
## age < 73.5 to the right, improve=1.8355900, (0 missing)
## depression < 0.5 to the left, improve=0.6554201, (0 missing)
## copd < 0.5 to the left, improve=0.3380891, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.2757170, (0 missing)
##
## Node number 15405: 21 observations
## predicted class=B1 expected loss=0.5238095 P(node) =7.641838e-05
## class counts: 10 2 6 3 0
## probabilities: 0.476 0.095 0.286 0.143 0.000
##
## Node number 15446: 1527 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5854617 P(node) =0.005556708
## class counts: 613 633 203 72 6
## probabilities: 0.401 0.415 0.133 0.047 0.004
## left son=30892 (1478 obs) right son=30893 (49 obs)
## Primary splits:
## reimbursement2008 < 3465 to the right, improve=1.7561930, (0 missing)
## age < 59.5 to the right, improve=1.2446620, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.3864080, (0 missing)
## stroke < 0.5 to the left, improve=0.3262151, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1237742, (0 missing)
##
## Node number 15447: 40 observations
## predicted class=B1 expected loss=0.65 P(node) =0.0001455588
## class counts: 14 12 12 2 0
## probabilities: 0.350 0.300 0.300 0.050 0.000
##
## Node number 15450: 703 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.598862 P(node) =0.002558196
## class counts: 282 244 119 53 5
## probabilities: 0.401 0.347 0.169 0.075 0.007
## left son=30900 (298 obs) right son=30901 (405 obs)
## Primary splits:
## reimbursement2008 < 6635 to the right, improve=1.9072210, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4867600, (0 missing)
## age < 74.5 to the left, improve=1.1374550, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4608058, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4586126, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.596, adj=0.047, (0 split)
## age < 35.5 to the left, agree=0.578, adj=0.003, (0 split)
##
## Node number 15451: 34 observations
## predicted class=B2 expected loss=0.5 P(node) =0.000123725
## class counts: 8 17 4 5 0
## probabilities: 0.235 0.500 0.118 0.147 0.000
##
## Node number 15452: 297 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5757576 P(node) =0.001080774
## class counts: 126 113 45 12 1
## probabilities: 0.424 0.380 0.152 0.040 0.003
## left son=30904 (274 obs) right son=30905 (23 obs)
## Primary splits:
## reimbursement2008 < 5065 to the left, improve=2.3768610, (0 missing)
## alzheimers < 0.5 to the right, improve=2.2936150, (0 missing)
## age < 37.5 to the left, improve=1.9456180, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.7719678, (0 missing)
## stroke < 0.5 to the left, improve=0.6098027, (0 missing)
##
## Node number 15453: 700 observations
## predicted class=B2 expected loss=0.5814286 P(node) =0.002547279
## class counts: 231 293 126 45 5
## probabilities: 0.330 0.419 0.180 0.064 0.007
##
## Node number 15458: 889 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.003235045
## class counts: 327 381 133 45 3
## probabilities: 0.368 0.429 0.150 0.051 0.003
##
## Node number 15459: 1053 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5318139 P(node) =0.003831836
## class counts: 315 493 175 64 6
## probabilities: 0.299 0.468 0.166 0.061 0.006
## left son=30918 (721 obs) right son=30919 (332 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.319421, (0 missing)
## age < 65.5 to the right, improve=2.157808, (0 missing)
## reimbursement2008 < 4195 to the left, improve=2.005955, (0 missing)
## stroke < 0.5 to the left, improve=1.694776, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.685891, (0 missing)
##
## Node number 15480: 83 observations
## predicted class=B2 expected loss=0.5662651 P(node) =0.0003020345
## class counts: 29 36 8 10 0
## probabilities: 0.349 0.434 0.096 0.120 0.000
##
## Node number 15481: 556 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6330935 P(node) =0.002023268
## class counts: 204 185 116 49 2
## probabilities: 0.367 0.333 0.209 0.088 0.004
## left son=30962 (368 obs) right son=30963 (188 obs)
## Primary splits:
## age < 67.5 to the right, improve=1.7538220, (0 missing)
## reimbursement2008 < 17290 to the right, improve=1.5233210, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8892958, (0 missing)
## stroke < 0.5 to the left, improve=0.8663588, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8033839, (0 missing)
##
## Node number 15484: 40 observations
## predicted class=B1 expected loss=0.425 P(node) =0.0001455588
## class counts: 23 8 7 2 0
## probabilities: 0.575 0.200 0.175 0.050 0.000
##
## Node number 15485: 82 observations
## predicted class=B2 expected loss=0.6219512 P(node) =0.0002983956
## class counts: 26 31 15 10 0
## probabilities: 0.317 0.378 0.183 0.122 0.000
##
## Node number 15884: 93 observations, complexity param=5.811572e-05
## predicted class=B2 expected loss=0.5913978 P(node) =0.0003384243
## class counts: 32 38 15 8 0
## probabilities: 0.344 0.409 0.161 0.086 0.000
## left son=31768 (44 obs) right son=31769 (49 obs)
## Primary splits:
## reimbursement2008 < 6110 to the right, improve=2.8038180, (0 missing)
## age < 68.5 to the right, improve=0.9063337, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4118188, (0 missing)
## alzheimers < 0.5 to the left, improve=0.3578690, (0 missing)
## stroke < 0.5 to the right, improve=0.3151562, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.785, adj=0.545, (0 split)
## age < 73.5 to the left, agree=0.602, adj=0.159, (0 split)
## alzheimers < 0.5 to the right, agree=0.538, adj=0.023, (0 split)
##
## Node number 15885: 13 observations
## predicted class=B1 expected loss=0.4615385 P(node) =4.730662e-05
## class counts: 7 1 1 4 0
## probabilities: 0.538 0.077 0.077 0.308 0.000
##
## Node number 15920: 331 observations
## predicted class=B1 expected loss=0.6344411 P(node) =0.001204499
## class counts: 121 94 53 53 10
## probabilities: 0.366 0.284 0.160 0.160 0.030
##
## Node number 15921: 18 observations
## predicted class=B2 expected loss=0.5 P(node) =6.550147e-05
## class counts: 2 9 2 4 1
## probabilities: 0.111 0.500 0.111 0.222 0.056
##
## Node number 15924: 310 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6741935 P(node) =0.001128081
## class counts: 101 99 64 37 9
## probabilities: 0.326 0.319 0.206 0.119 0.029
## left son=31848 (50 obs) right son=31849 (260 obs)
## Primary splits:
## reimbursement2008 < 9955 to the left, improve=3.5194040, (0 missing)
## alzheimers < 0.5 to the left, improve=1.4052180, (0 missing)
## age < 60.5 to the right, improve=1.3545900, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9417634, (0 missing)
## stroke < 0.5 to the right, improve=0.4401818, (0 missing)
##
## Node number 15925: 237 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6582278 P(node) =0.000862436
## class counts: 53 81 72 28 3
## probabilities: 0.224 0.342 0.304 0.118 0.013
## left son=31850 (56 obs) right son=31851 (181 obs)
## Primary splits:
## age < 67.5 to the left, improve=3.14488400, (0 missing)
## reimbursement2008 < 7130 to the left, improve=2.11196700, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.86604090, (0 missing)
## stroke < 0.5 to the left, improve=0.39390990, (0 missing)
## alzheimers < 0.5 to the right, improve=0.05008339, (0 missing)
##
## Node number 16280: 398 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7236181 P(node) =0.00144831
## class counts: 110 101 63 99 25
## probabilities: 0.276 0.254 0.158 0.249 0.063
## left son=32560 (179 obs) right son=32561 (219 obs)
## Primary splits:
## alzheimers < 0.5 to the right, improve=2.797541, (0 missing)
## ihd < 0.5 to the right, improve=2.182276, (0 missing)
## reimbursement2008 < 15500 to the right, improve=1.710577, (0 missing)
## stroke < 0.5 to the right, improve=1.223226, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.211249, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the right, agree=0.563, adj=0.028, (0 split)
## reimbursement2008 < 15625 to the left, agree=0.563, adj=0.028, (0 split)
## age < 62.5 to the left, agree=0.555, adj=0.011, (0 split)
## osteoporosis < 0.5 to the right, agree=0.553, adj=0.006, (0 split)
##
## Node number 16281: 59 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6440678 P(node) =0.0002146993
## class counts: 21 6 10 21 1
## probabilities: 0.356 0.102 0.169 0.356 0.017
## left son=32562 (18 obs) right son=32563 (41 obs)
## Primary splits:
## reimbursement2008 < 19680 to the right, improve=1.9754260, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9501021, (0 missing)
## bucket2008 < 3.5 to the right, improve=0.8931654, (0 missing)
## age < 90.5 to the right, improve=0.7250257, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5260164, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.847, adj=0.5, (0 split)
##
## Node number 27764: 343 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5568513 P(node) =0.001248167
## class counts: 152 136 37 16 2
## probabilities: 0.443 0.397 0.108 0.047 0.006
## left son=55528 (117 obs) right son=55529 (226 obs)
## Primary splits:
## reimbursement2008 < 2835 to the right, improve=1.9282960, (0 missing)
## stroke < 0.5 to the left, improve=1.1581140, (0 missing)
## age < 82.5 to the right, improve=1.0933820, (0 missing)
## kidney < 0.5 to the left, improve=1.0145490, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9380155, (0 missing)
##
## Node number 27765: 129 observations
## predicted class=B1 expected loss=0.4496124 P(node) =0.0004694272
## class counts: 71 27 20 9 2
## probabilities: 0.550 0.209 0.155 0.070 0.016
##
## Node number 27766: 218 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5733945 P(node) =0.0007932956
## class counts: 93 89 28 8 0
## probabilities: 0.427 0.408 0.128 0.037 0.000
## left son=55532 (194 obs) right son=55533 (24 obs)
## Primary splits:
## reimbursement2008 < 2945 to the left, improve=1.9617420, (0 missing)
## depression < 0.5 to the left, improve=0.6526821, (0 missing)
## kidney < 0.5 to the left, improve=0.4610298, (0 missing)
## age < 57.5 to the left, improve=0.4574831, (0 missing)
## alzheimers < 0.5 to the left, improve=0.3559027, (0 missing)
##
## Node number 27767: 79 observations
## predicted class=B2 expected loss=0.4810127 P(node) =0.0002874787
## class counts: 26 41 7 5 0
## probabilities: 0.329 0.519 0.089 0.063 0.000
##
## Node number 28002: 192 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.59375 P(node) =0.0006986823
## class counts: 72 78 29 11 2
## probabilities: 0.375 0.406 0.151 0.057 0.010
## left son=56004 (124 obs) right son=56005 (68 obs)
## Primary splits:
## age < 78.5 to the left, improve=3.1968100, (0 missing)
## reimbursement2008 < 2885 to the left, improve=2.1236740, (0 missing)
## alzheimers < 0.5 to the right, improve=1.0053880, (0 missing)
## bucket2008 < 1.5 to the left, improve=0.7479369, (0 missing)
## stroke < 0.5 to the left, improve=0.5316513, (0 missing)
##
## Node number 28003: 74 observations, complexity param=7.19528e-05
## predicted class=B3 expected loss=0.6486486 P(node) =0.0002692838
## class counts: 24 22 26 2 0
## probabilities: 0.324 0.297 0.351 0.027 0.000
## left son=56006 (8 obs) right son=56007 (66 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=6.4864860, (0 missing)
## age < 65 to the left, improve=1.7666590, (0 missing)
## alzheimers < 0.5 to the left, improve=1.7622440, (0 missing)
## reimbursement2008 < 2355 to the right, improve=1.0927360, (0 missing)
## stroke < 0.5 to the left, improve=0.8745462, (0 missing)
##
## Node number 28066: 169 observations
## predicted class=B1 expected loss=0.5798817 P(node) =0.000614986
## class counts: 71 58 32 6 2
## probabilities: 0.420 0.343 0.189 0.036 0.012
##
## Node number 28067: 45 observations
## predicted class=B2 expected loss=0.5111111 P(node) =0.0001637537
## class counts: 12 22 9 2 0
## probabilities: 0.267 0.489 0.200 0.044 0.000
##
## Node number 28354: 169 observations
## predicted class=B1 expected loss=0.5621302 P(node) =0.000614986
## class counts: 74 60 26 8 1
## probabilities: 0.438 0.355 0.154 0.047 0.006
##
## Node number 28355: 62 observations
## predicted class=B2 expected loss=0.4516129 P(node) =0.0002256162
## class counts: 20 34 6 2 0
## probabilities: 0.323 0.548 0.097 0.032 0.000
##
## Node number 28358: 689 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5718433 P(node) =0.002507251
## class counts: 295 261 92 38 3
## probabilities: 0.428 0.379 0.134 0.055 0.004
## left son=56716 (367 obs) right son=56717 (322 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=1.5366830, (0 missing)
## reimbursement2008 < 2185 to the right, improve=0.9498001, (0 missing)
## age < 67.5 to the left, improve=0.9450906, (0 missing)
## copd < 0.5 to the left, improve=0.5052370, (0 missing)
## depression < 0.5 to the left, improve=0.4336301, (0 missing)
## Surrogate splits:
## copd < 0.5 to the left, agree=0.605, adj=0.155, (0 split)
## age < 85.5 to the left, agree=0.543, adj=0.022, (0 split)
## reimbursement2008 < 2515 to the left, agree=0.541, adj=0.019, (0 split)
## alzheimers < 0.5 to the left, agree=0.538, adj=0.012, (0 split)
##
## Node number 28359: 34 observations
## predicted class=B2 expected loss=0.5294118 P(node) =0.000123725
## class counts: 8 16 6 4 0
## probabilities: 0.235 0.471 0.176 0.118 0.000
##
## Node number 30808: 253 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5454545 P(node) =0.0009206595
## class counts: 115 89 30 17 2
## probabilities: 0.455 0.352 0.119 0.067 0.008
## left son=61616 (245 obs) right son=61617 (8 obs)
## Primary splits:
## age < 96 to the left, improve=1.6668230, (0 missing)
## reimbursement2008 < 8170 to the left, improve=1.5801570, (0 missing)
## depression < 0.5 to the left, improve=0.8012407, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6406559, (0 missing)
## stroke < 0.5 to the left, improve=0.4810539, (0 missing)
##
## Node number 30809: 56 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0002037823
## class counts: 17 28 8 3 0
## probabilities: 0.304 0.500 0.143 0.054 0.000
##
## Node number 30892: 1478 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5886333 P(node) =0.005378398
## class counts: 600 608 199 67 4
## probabilities: 0.406 0.411 0.135 0.045 0.003
## left son=61784 (759 obs) right son=61785 (719 obs)
## Primary splits:
## reimbursement2008 < 4655 to the left, improve=1.4912330, (0 missing)
## age < 59.5 to the right, improve=1.4379920, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4252592, (0 missing)
## stroke < 0.5 to the left, improve=0.4189515, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.1287486, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.566, adj=0.108, (0 split)
## alzheimers < 0.5 to the left, agree=0.535, adj=0.043, (0 split)
## stroke < 0.5 to the left, agree=0.533, adj=0.040, (0 split)
## copd < 0.5 to the left, agree=0.530, adj=0.033, (0 split)
## age < 82.5 to the left, agree=0.526, adj=0.025, (0 split)
##
## Node number 30893: 49 observations
## predicted class=B2 expected loss=0.4897959 P(node) =0.0001783096
## class counts: 13 25 4 5 2
## probabilities: 0.265 0.510 0.082 0.102 0.041
##
## Node number 30900: 298 observations
## predicted class=B1 expected loss=0.5503356 P(node) =0.001084413
## class counts: 134 94 46 20 4
## probabilities: 0.450 0.315 0.154 0.067 0.013
##
## Node number 30901: 405 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.6296296 P(node) =0.001473783
## class counts: 148 150 73 33 1
## probabilities: 0.365 0.370 0.180 0.081 0.002
## left son=61802 (137 obs) right son=61803 (268 obs)
## Primary splits:
## reimbursement2008 < 5685 to the left, improve=1.4352860, (0 missing)
## age < 43 to the right, improve=1.2563810, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8108852, (0 missing)
## alzheimers < 0.5 to the right, improve=0.3644866, (0 missing)
## copd < 0.5 to the left, improve=0.3421456, (0 missing)
##
## Node number 30904: 274 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5583942 P(node) =0.0009970779
## class counts: 121 99 42 11 1
## probabilities: 0.442 0.361 0.153 0.040 0.004
## left son=61808 (174 obs) right son=61809 (100 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.2349370, (0 missing)
## age < 37.5 to the left, improve=1.7714310, (0 missing)
## reimbursement2008 < 4990 to the right, improve=1.7636660, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8701440, (0 missing)
## stroke < 0.5 to the left, improve=0.4025273, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.668, adj=0.09, (0 split)
## reimbursement2008 < 3085 to the right, agree=0.642, adj=0.02, (0 split)
##
## Node number 30905: 23 observations
## predicted class=B2 expected loss=0.3913043 P(node) =8.369632e-05
## class counts: 5 14 3 1 0
## probabilities: 0.217 0.609 0.130 0.043 0.000
##
## Node number 30918: 721 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5492372 P(node) =0.002623698
## class counts: 234 325 114 44 4
## probabilities: 0.325 0.451 0.158 0.061 0.006
## left son=61836 (109 obs) right son=61837 (612 obs)
## Primary splits:
## age < 86.5 to the right, improve=5.2024390, (0 missing)
## reimbursement2008 < 8105 to the left, improve=1.9497410, (0 missing)
## stroke < 0.5 to the left, improve=1.3441110, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.8592657, (0 missing)
## alzheimers < 0.5 to the left, improve=0.1415473, (0 missing)
##
## Node number 30919: 332 observations
## predicted class=B2 expected loss=0.4939759 P(node) =0.001208138
## class counts: 81 168 61 20 2
## probabilities: 0.244 0.506 0.184 0.060 0.006
##
## Node number 30962: 368 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.625 P(node) =0.001339141
## class counts: 138 132 66 32 0
## probabilities: 0.375 0.359 0.179 0.087 0.000
## left son=61924 (261 obs) right son=61925 (107 obs)
## Primary splits:
## reimbursement2008 < 10440 to the right, improve=2.0386870, (0 missing)
## age < 68.5 to the right, improve=2.0238320, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.0604210, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8507150, (0 missing)
## heart.failure < 0.5 to the right, improve=0.3195541, (0 missing)
##
## Node number 30963: 188 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6489362 P(node) =0.0006841264
## class counts: 66 53 50 17 2
## probabilities: 0.351 0.282 0.266 0.090 0.011
## left son=61926 (135 obs) right son=61927 (53 obs)
## Primary splits:
## age < 55.5 to the right, improve=1.3142350, (0 missing)
## reimbursement2008 < 8995 to the left, improve=1.1323620, (0 missing)
## stroke < 0.5 to the left, improve=0.7672950, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7658279, (0 missing)
## bucket2008 < 3.5 to the right, improve=0.3998270, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8645 to the right, agree=0.723, adj=0.019, (0 split)
##
## Node number 31768: 44 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5227273 P(node) =0.0001601147
## class counts: 21 13 5 5 0
## probabilities: 0.477 0.295 0.114 0.114 0.000
## left son=63536 (26 obs) right son=63537 (18 obs)
## Primary splits:
## reimbursement2008 < 9180 to the left, improve=3.34188000, (0 missing)
## age < 73.5 to the right, improve=1.53473700, (0 missing)
## bucket2008 < 2.5 to the left, improve=1.08333300, (0 missing)
## alzheimers < 0.5 to the left, improve=0.99564270, (0 missing)
## copd < 0.5 to the right, improve=0.09090909, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.864, adj=0.667, (0 split)
## age < 82.5 to the left, agree=0.636, adj=0.111, (0 split)
## stroke < 0.5 to the left, agree=0.614, adj=0.056, (0 split)
##
## Node number 31769: 49 observations
## predicted class=B2 expected loss=0.4897959 P(node) =0.0001783096
## class counts: 11 25 10 3 0
## probabilities: 0.224 0.510 0.204 0.061 0.000
##
## Node number 31848: 50 observations
## predicted class=B2 expected loss=0.56 P(node) =0.0001819485
## class counts: 21 22 1 6 0
## probabilities: 0.420 0.440 0.020 0.120 0.000
##
## Node number 31849: 260 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6923077 P(node) =0.0009461323
## class counts: 80 77 63 31 9
## probabilities: 0.308 0.296 0.242 0.119 0.035
## left son=63698 (20 obs) right son=63699 (240 obs)
## Primary splits:
## reimbursement2008 < 14765 to the right, improve=1.866026, (0 missing)
## alzheimers < 0.5 to the left, improve=1.724179, (0 missing)
## age < 59 to the right, improve=1.389622, (0 missing)
## heart.failure < 0.5 to the left, improve=1.186623, (0 missing)
## stroke < 0.5 to the right, improve=0.396978, (0 missing)
##
## Node number 31850: 56 observations
## predicted class=B2 expected loss=0.5178571 P(node) =0.0002037823
## class counts: 11 27 9 9 0
## probabilities: 0.196 0.482 0.161 0.161 0.000
##
## Node number 31851: 181 observations, complexity param=5.534831e-05
## predicted class=B3 expected loss=0.6519337 P(node) =0.0006586537
## class counts: 42 54 63 19 3
## probabilities: 0.232 0.298 0.348 0.105 0.017
## left son=63702 (136 obs) right son=63703 (45 obs)
## Primary splits:
## reimbursement2008 < 6865 to the right, improve=2.6510090, (0 missing)
## age < 95.5 to the left, improve=1.1712710, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.4758931, (0 missing)
## stroke < 0.5 to the left, improve=0.1841866, (0 missing)
## heart.failure < 0.5 to the left, improve=0.1010412, (0 missing)
##
## Node number 32560: 179 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.6815642 P(node) =0.0006513757
## class counts: 57 51 27 31 13
## probabilities: 0.318 0.285 0.151 0.173 0.073
## left son=65120 (38 obs) right son=65121 (141 obs)
## Primary splits:
## reimbursement2008 < 21440 to the right, improve=2.8400160, (0 missing)
## age < 70.5 to the left, improve=1.0471050, (0 missing)
## stroke < 0.5 to the right, improve=0.8887163, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8119666, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6975642, (0 missing)
##
## Node number 32561: 219 observations
## predicted class=B4 expected loss=0.6894977 P(node) =0.0007969345
## class counts: 53 50 36 68 12
## probabilities: 0.242 0.228 0.164 0.311 0.055
##
## Node number 32562: 18 observations
## predicted class=B1 expected loss=0.4444444 P(node) =6.550147e-05
## class counts: 10 2 0 5 1
## probabilities: 0.556 0.111 0.000 0.278 0.056
##
## Node number 32563: 41 observations
## predicted class=B4 expected loss=0.6097561 P(node) =0.0001491978
## class counts: 11 4 10 16 0
## probabilities: 0.268 0.098 0.244 0.390 0.000
##
## Node number 55528: 117 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4871795 P(node) =0.0004257595
## class counts: 60 38 15 3 1
## probabilities: 0.513 0.325 0.128 0.026 0.009
## left son=111056 (78 obs) right son=111057 (39 obs)
## Primary splits:
## reimbursement2008 < 2945 to the left, improve=2.9829060, (0 missing)
## kidney < 0.5 to the left, improve=1.2210830, (0 missing)
## age < 76.5 to the right, improve=1.1210830, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9103070, (0 missing)
## stroke < 0.5 to the left, improve=0.2543679, (0 missing)
## Surrogate splits:
## age < 76.5 to the right, agree=0.684, adj=0.051, (0 split)
##
## Node number 55529: 226 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5663717 P(node) =0.0008224073
## class counts: 92 98 22 13 1
## probabilities: 0.407 0.434 0.097 0.058 0.004
## left son=111058 (72 obs) right son=111059 (154 obs)
## Primary splits:
## age < 80.5 to the right, improve=1.5914710, (0 missing)
## stroke < 0.5 to the left, improve=1.3555880, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7668454, (0 missing)
## reimbursement2008 < 2795 to the left, improve=0.6874895, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4980787, (0 missing)
##
## Node number 55532: 194 observations
## predicted class=B1 expected loss=0.556701 P(node) =0.0007059603
## class counts: 86 74 27 7 0
## probabilities: 0.443 0.381 0.139 0.036 0.000
##
## Node number 55533: 24 observations
## predicted class=B2 expected loss=0.375 P(node) =8.733529e-05
## class counts: 7 15 1 1 0
## probabilities: 0.292 0.625 0.042 0.042 0.000
##
## Node number 56004: 124 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5403226 P(node) =0.0004512323
## class counts: 57 47 16 4 0
## probabilities: 0.460 0.379 0.129 0.032 0.000
## left son=112008 (46 obs) right son=112009 (78 obs)
## Primary splits:
## age < 72.5 to the right, improve=2.8817380, (0 missing)
## reimbursement2008 < 2885 to the left, improve=1.5254660, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4454760, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7103829, (0 missing)
## stroke < 0.5 to the left, improve=0.4023915, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the right, agree=0.661, adj=0.087, (0 split)
## reimbursement2008 < 2575 to the left, agree=0.645, adj=0.043, (0 split)
##
## Node number 56005: 68 observations
## predicted class=B2 expected loss=0.5441176 P(node) =0.00024745
## class counts: 15 31 13 7 2
## probabilities: 0.221 0.456 0.191 0.103 0.029
##
## Node number 56006: 8 observations
## predicted class=B2 expected loss=0 P(node) =2.911176e-05
## class counts: 0 8 0 0 0
## probabilities: 0.000 1.000 0.000 0.000 0.000
##
## Node number 56007: 66 observations, complexity param=5.534831e-05
## predicted class=B3 expected loss=0.6060606 P(node) =0.0002401721
## class counts: 24 14 26 2 0
## probabilities: 0.364 0.212 0.394 0.030 0.000
## left son=112014 (40 obs) right son=112015 (26 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.3576920, (0 missing)
## age < 65 to the left, improve=1.8352940, (0 missing)
## reimbursement2008 < 2375 to the right, improve=1.1494920, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.1363636, (0 missing)
## Surrogate splits:
## age < 82.5 to the left, agree=0.652, adj=0.115, (0 split)
##
## Node number 56716: 367 observations
## predicted class=B1 expected loss=0.5395095 P(node) =0.001335502
## class counts: 169 131 51 13 3
## probabilities: 0.460 0.357 0.139 0.035 0.008
##
## Node number 56717: 322 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5962733 P(node) =0.001171748
## class counts: 126 130 41 25 0
## probabilities: 0.391 0.404 0.127 0.078 0.000
## left son=113434 (78 obs) right son=113435 (244 obs)
## Primary splits:
## age < 67.5 to the left, improve=2.0124890, (0 missing)
## reimbursement2008 < 2265 to the right, improve=1.1949400, (0 missing)
## alzheimers < 0.5 to the right, improve=0.3273471, (0 missing)
## depression < 0.5 to the right, improve=0.1786959, (0 missing)
## copd < 0.5 to the left, improve=0.1745923, (0 missing)
##
## Node number 61616: 245 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5346939 P(node) =0.0008915478
## class counts: 114 87 27 16 1
## probabilities: 0.465 0.355 0.110 0.065 0.004
## left son=123232 (209 obs) right son=123233 (36 obs)
## Primary splits:
## reimbursement2008 < 8170 to the left, improve=1.7182870, (0 missing)
## age < 90.5 to the right, improve=1.6062760, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.7459219, (0 missing)
## depression < 0.5 to the left, improve=0.6596720, (0 missing)
## stroke < 0.5 to the left, improve=0.6366849, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.971, adj=0.806, (0 split)
##
## Node number 61617: 8 observations
## predicted class=B3 expected loss=0.625 P(node) =2.911176e-05
## class counts: 1 2 3 1 1
## probabilities: 0.125 0.250 0.375 0.125 0.125
##
## Node number 61784: 759 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5678524 P(node) =0.002761979
## class counts: 328 303 94 33 1
## probabilities: 0.432 0.399 0.124 0.043 0.001
## left son=123568 (158 obs) right son=123569 (601 obs)
## Primary splits:
## reimbursement2008 < 4315 to the right, improve=1.62186500, (0 missing)
## age < 82.5 to the right, improve=0.60286370, (0 missing)
## alzheimers < 0.5 to the right, improve=0.24697950, (0 missing)
## copd < 0.5 to the left, improve=0.10233690, (0 missing)
## stroke < 0.5 to the left, improve=0.09394217, (0 missing)
##
## Node number 61785: 719 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5757997 P(node) =0.00261642
## class counts: 272 305 105 34 3
## probabilities: 0.378 0.424 0.146 0.047 0.004
## left son=123570 (346 obs) right son=123571 (373 obs)
## Primary splits:
## reimbursement2008 < 5835 to the left, improve=2.8015510, (0 missing)
## age < 59.5 to the right, improve=2.2849680, (0 missing)
## alzheimers < 0.5 to the right, improve=0.5855315, (0 missing)
## stroke < 0.5 to the right, improve=0.5109046, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2469968, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.590, adj=0.147, (0 split)
## alzheimers < 0.5 to the left, agree=0.537, adj=0.038, (0 split)
## age < 60.5 to the left, agree=0.527, adj=0.017, (0 split)
##
## Node number 61802: 137 observations
## predicted class=B1 expected loss=0.5839416 P(node) =0.000498539
## class counts: 57 43 22 15 0
## probabilities: 0.416 0.314 0.161 0.109 0.000
##
## Node number 61803: 268 observations
## predicted class=B2 expected loss=0.6007463 P(node) =0.0009752441
## class counts: 91 107 51 18 1
## probabilities: 0.340 0.399 0.190 0.067 0.004
##
## Node number 61808: 174 observations
## predicted class=B1 expected loss=0.5229885 P(node) =0.0006331809
## class counts: 83 53 29 8 1
## probabilities: 0.477 0.305 0.167 0.046 0.006
##
## Node number 61809: 100 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.54 P(node) =0.000363897
## class counts: 38 46 13 3 0
## probabilities: 0.380 0.460 0.130 0.030 0.000
## left son=123618 (26 obs) right son=123619 (74 obs)
## Primary splits:
## reimbursement2008 < 4355 to the right, improve=5.3372560, (0 missing)
## age < 62.5 to the right, improve=1.9704690, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.2703500, (0 missing)
## stroke < 0.5 to the left, improve=0.5886275, (0 missing)
## Surrogate splits:
## age < 50.5 to the left, agree=0.79, adj=0.192, (0 split)
##
## Node number 61836: 109 observations
## predicted class=B1 expected loss=0.5412844 P(node) =0.0003966478
## class counts: 50 33 16 9 1
## probabilities: 0.459 0.303 0.147 0.083 0.009
##
## Node number 61837: 612 observations
## predicted class=B2 expected loss=0.5228758 P(node) =0.00222705
## class counts: 184 292 98 35 3
## probabilities: 0.301 0.477 0.160 0.057 0.005
##
## Node number 61924: 261 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5862069 P(node) =0.0009497713
## class counts: 108 87 44 22 0
## probabilities: 0.414 0.333 0.169 0.084 0.000
## left son=123848 (92 obs) right son=123849 (169 obs)
## Primary splits:
## reimbursement2008 < 12585 to the left, improve=2.1315740, (0 missing)
## age < 77.5 to the right, improve=1.2761660, (0 missing)
## stroke < 0.5 to the left, improve=1.0543160, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.0296720, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5202291, (0 missing)
##
## Node number 61925: 107 observations
## predicted class=B2 expected loss=0.5794393 P(node) =0.0003893698
## class counts: 30 45 22 10 0
## probabilities: 0.280 0.421 0.206 0.093 0.000
##
## Node number 61926: 135 observations
## predicted class=B1 expected loss=0.6074074 P(node) =0.000491261
## class counts: 53 34 36 12 0
## probabilities: 0.393 0.252 0.267 0.089 0.000
##
## Node number 61927: 53 observations
## predicted class=B2 expected loss=0.6415094 P(node) =0.0001928654
## class counts: 13 19 14 5 2
## probabilities: 0.245 0.358 0.264 0.094 0.038
##
## Node number 63536: 26 observations
## predicted class=B1 expected loss=0.3461538 P(node) =9.461323e-05
## class counts: 17 4 2 3 0
## probabilities: 0.654 0.154 0.077 0.115 0.000
##
## Node number 63537: 18 observations
## predicted class=B2 expected loss=0.5 P(node) =6.550147e-05
## class counts: 4 9 3 2 0
## probabilities: 0.222 0.500 0.167 0.111 0.000
##
## Node number 63698: 20 observations
## predicted class=B1 expected loss=0.45 P(node) =7.277941e-05
## class counts: 11 5 2 1 1
## probabilities: 0.550 0.250 0.100 0.050 0.050
##
## Node number 63699: 240 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.7 P(node) =0.0008733529
## class counts: 69 72 61 30 8
## probabilities: 0.288 0.300 0.254 0.125 0.033
## left son=127398 (201 obs) right son=127399 (39 obs)
## Primary splits:
## age < 61.5 to the right, improve=1.4580460, (0 missing)
## reimbursement2008 < 10970 to the right, improve=1.4206140, (0 missing)
## alzheimers < 0.5 to the left, improve=1.0755290, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9752886, (0 missing)
## stroke < 0.5 to the right, improve=0.4524283, (0 missing)
##
## Node number 63702: 136 observations
## predicted class=B3 expected loss=0.6029412 P(node) =0.0004949
## class counts: 34 36 54 10 2
## probabilities: 0.250 0.265 0.397 0.074 0.015
##
## Node number 63703: 45 observations
## predicted class=B2 expected loss=0.6 P(node) =0.0001637537
## class counts: 8 18 9 9 1
## probabilities: 0.178 0.400 0.200 0.200 0.022
##
## Node number 65120: 38 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0001382809
## class counts: 9 19 4 5 1
## probabilities: 0.237 0.500 0.105 0.132 0.026
##
## Node number 65121: 141 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6595745 P(node) =0.0005130948
## class counts: 48 32 23 26 12
## probabilities: 0.340 0.227 0.163 0.184 0.085
## left son=130242 (89 obs) right son=130243 (52 obs)
## Primary splits:
## reimbursement2008 < 17585 to the right, improve=2.1889060, (0 missing)
## age < 47.5 to the right, improve=1.2186760, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.1702130, (0 missing)
## stroke < 0.5 to the right, improve=0.9175166, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5919705, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.702, adj=0.192, (0 split)
## age < 47.5 to the right, agree=0.667, adj=0.096, (0 split)
##
## Node number 111056: 78 observations
## predicted class=B1 expected loss=0.4102564 P(node) =0.0002838397
## class counts: 46 19 11 2 0
## probabilities: 0.590 0.244 0.141 0.026 0.000
##
## Node number 111057: 39 observations
## predicted class=B2 expected loss=0.5128205 P(node) =0.0001419198
## class counts: 14 19 4 1 1
## probabilities: 0.359 0.487 0.103 0.026 0.026
##
## Node number 111058: 72 observations
## predicted class=B1 expected loss=0.4861111 P(node) =0.0002620059
## class counts: 37 29 5 1 0
## probabilities: 0.514 0.403 0.069 0.014 0.000
##
## Node number 111059: 154 observations
## predicted class=B2 expected loss=0.5519481 P(node) =0.0005604015
## class counts: 55 69 17 12 1
## probabilities: 0.357 0.448 0.110 0.078 0.006
##
## Node number 112008: 46 observations
## predicted class=B1 expected loss=0.4347826 P(node) =0.0001673926
## class counts: 26 10 8 2 0
## probabilities: 0.565 0.217 0.174 0.043 0.000
##
## Node number 112009: 78 observations
## predicted class=B2 expected loss=0.525641 P(node) =0.0002838397
## class counts: 31 37 8 2 0
## probabilities: 0.397 0.474 0.103 0.026 0.000
##
## Node number 112014: 40 observations
## predicted class=B1 expected loss=0.6 P(node) =0.0001455588
## class counts: 16 12 11 1 0
## probabilities: 0.400 0.300 0.275 0.025 0.000
##
## Node number 112015: 26 observations
## predicted class=B3 expected loss=0.4230769 P(node) =9.461323e-05
## class counts: 8 2 15 1 0
## probabilities: 0.308 0.077 0.577 0.038 0.000
##
## Node number 113434: 78 observations
## predicted class=B1 expected loss=0.5512821 P(node) =0.0002838397
## class counts: 35 23 15 5 0
## probabilities: 0.449 0.295 0.192 0.064 0.000
##
## Node number 113435: 244 observations
## predicted class=B2 expected loss=0.5614754 P(node) =0.0008879088
## class counts: 91 107 26 20 0
## probabilities: 0.373 0.439 0.107 0.082 0.000
##
## Node number 123232: 209 observations
## predicted class=B1 expected loss=0.507177 P(node) =0.0007605448
## class counts: 103 70 22 14 0
## probabilities: 0.493 0.335 0.105 0.067 0.000
##
## Node number 123233: 36 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5277778 P(node) =0.0001310029
## class counts: 11 17 5 2 1
## probabilities: 0.306 0.472 0.139 0.056 0.028
## left son=246466 (15 obs) right son=246467 (21 obs)
## Primary splits:
## age < 74.5 to the left, improve=5.3968250, (0 missing)
## reimbursement2008 < 8705 to the right, improve=1.5053320, (0 missing)
## copd < 0.5 to the right, improve=0.3703704, (0 missing)
## depression < 0.5 to the right, improve=0.3527778, (0 missing)
## heart.failure < 0.5 to the left, improve=0.2972583, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8460 to the left, agree=0.611, adj=0.067, (0 split)
##
## Node number 123568: 158 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0005749573
## class counts: 79 55 15 8 1
## probabilities: 0.500 0.348 0.095 0.051 0.006
##
## Node number 123569: 601 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5856905 P(node) =0.002187021
## class counts: 249 248 79 25 0
## probabilities: 0.414 0.413 0.131 0.042 0.000
## left son=247138 (592 obs) right son=247139 (9 obs)
## Primary splits:
## reimbursement2008 < 4295 to the left, improve=3.0859230, (0 missing)
## age < 62.5 to the right, improve=0.8258296, (0 missing)
## copd < 0.5 to the left, improve=0.2730143, (0 missing)
## stroke < 0.5 to the left, improve=0.1209200, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1102521, (0 missing)
##
## Node number 123570: 346 observations
## predicted class=B2 expected loss=0.5202312 P(node) =0.001259084
## class counts: 122 166 44 13 1
## probabilities: 0.353 0.480 0.127 0.038 0.003
##
## Node number 123571: 373 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.5978552 P(node) =0.001357336
## class counts: 150 139 61 21 2
## probabilities: 0.402 0.373 0.164 0.056 0.005
## left son=247142 (124 obs) right son=247143 (249 obs)
## Primary splits:
## alzheimers < 0.5 to the right, improve=1.9370400, (0 missing)
## reimbursement2008 < 6045 to the right, improve=1.9317030, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.1351660, (0 missing)
## age < 68.5 to the left, improve=0.9923350, (0 missing)
## stroke < 0.5 to the right, improve=0.8206414, (0 missing)
## Surrogate splits:
## age < 64.5 to the left, agree=0.673, adj=0.016, (0 split)
## stroke < 0.5 to the right, agree=0.673, adj=0.016, (0 split)
## reimbursement2008 < 5845 to the left, agree=0.670, adj=0.008, (0 split)
##
## Node number 123618: 26 observations
## predicted class=B1 expected loss=0.3846154 P(node) =9.461323e-05
## class counts: 16 4 4 2 0
## probabilities: 0.615 0.154 0.154 0.077 0.000
##
## Node number 123619: 74 observations
## predicted class=B2 expected loss=0.4324324 P(node) =0.0002692838
## class counts: 22 42 9 1 0
## probabilities: 0.297 0.568 0.122 0.014 0.000
##
## Node number 123848: 92 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0003347853
## class counts: 46 23 17 6 0
## probabilities: 0.500 0.250 0.185 0.065 0.000
##
## Node number 123849: 169 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6213018 P(node) =0.000614986
## class counts: 62 64 27 16 0
## probabilities: 0.367 0.379 0.160 0.095 0.000
## left son=247698 (109 obs) right son=247699 (60 obs)
## Primary splits:
## reimbursement2008 < 14485 to the right, improve=2.3703890, (0 missing)
## age < 77.5 to the right, improve=1.8205180, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.5605270, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9473954, (0 missing)
## stroke < 0.5 to the right, improve=0.8779250, (0 missing)
##
## Node number 127398: 201 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6865672 P(node) =0.0007314331
## class counts: 63 58 49 27 4
## probabilities: 0.313 0.289 0.244 0.134 0.020
## left son=254796 (112 obs) right son=254797 (89 obs)
## Primary splits:
## reimbursement2008 < 12625 to the left, improve=2.6465710, (0 missing)
## age < 72.5 to the right, improve=1.5701210, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5204340, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8281641, (0 missing)
## stroke < 0.5 to the right, improve=0.4454147, (0 missing)
## Surrogate splits:
## age < 67.5 to the right, agree=0.587, adj=0.067, (0 split)
##
## Node number 127399: 39 observations
## predicted class=B2 expected loss=0.6410256 P(node) =0.0001419198
## class counts: 6 14 12 3 4
## probabilities: 0.154 0.359 0.308 0.077 0.103
##
## Node number 130242: 89 observations
## predicted class=B1 expected loss=0.5955056 P(node) =0.0003238684
## class counts: 36 15 17 14 7
## probabilities: 0.404 0.169 0.191 0.157 0.079
##
## Node number 130243: 52 observations
## predicted class=B2 expected loss=0.6730769 P(node) =0.0001892265
## class counts: 12 17 6 12 5
## probabilities: 0.231 0.327 0.115 0.231 0.096
##
## Node number 246466: 15 observations
## predicted class=B1 expected loss=0.4 P(node) =5.458456e-05
## class counts: 9 2 3 0 1
## probabilities: 0.600 0.133 0.200 0.000 0.067
##
## Node number 246467: 21 observations
## predicted class=B2 expected loss=0.2857143 P(node) =7.641838e-05
## class counts: 2 15 2 2 0
## probabilities: 0.095 0.714 0.095 0.095 0.000
##
## Node number 247138: 592 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5810811 P(node) =0.002154271
## class counts: 248 240 79 25 0
## probabilities: 0.419 0.405 0.133 0.042 0.000
## left son=494276 (135 obs) right son=494277 (457 obs)
## Primary splits:
## age < 82.5 to the right, improve=1.0162580, (0 missing)
## reimbursement2008 < 3485 to the left, improve=0.9533819, (0 missing)
## copd < 0.5 to the left, improve=0.2603666, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1489946, (0 missing)
## stroke < 0.5 to the left, improve=0.1384892, (0 missing)
##
## Node number 247139: 9 observations
## predicted class=B2 expected loss=0.1111111 P(node) =3.275073e-05
## class counts: 1 8 0 0 0
## probabilities: 0.111 0.889 0.000 0.000 0.000
##
## Node number 247142: 124 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.516129 P(node) =0.0004512323
## class counts: 60 39 19 5 1
## probabilities: 0.484 0.315 0.153 0.040 0.008
## left son=494284 (114 obs) right son=494285 (10 obs)
## Primary splits:
## reimbursement2008 < 8555 to the left, improve=3.2894170, (0 missing)
## age < 62.5 to the right, improve=1.3134040, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8306452, (0 missing)
## stroke < 0.5 to the right, improve=0.6624062, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6169185, (0 missing)
##
## Node number 247143: 249 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5983936 P(node) =0.0009061036
## class counts: 90 100 42 16 1
## probabilities: 0.361 0.402 0.169 0.064 0.004
## left son=494286 (217 obs) right son=494287 (32 obs)
## Primary splits:
## reimbursement2008 < 6045 to the right, improve=2.8382200, (0 missing)
## age < 68.5 to the left, improve=1.5757780, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8580882, (0 missing)
## copd < 0.5 to the left, improve=0.4427711, (0 missing)
## stroke < 0.5 to the right, improve=0.2244234, (0 missing)
##
## Node number 247698: 109 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5779817 P(node) =0.0003966478
## class counts: 46 34 19 10 0
## probabilities: 0.422 0.312 0.174 0.092 0.000
## left son=495396 (72 obs) right son=495397 (37 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.1824740, (0 missing)
## reimbursement2008 < 17235 to the right, improve=1.3957040, (0 missing)
## age < 70.5 to the left, improve=1.2827700, (0 missing)
## stroke < 0.5 to the left, improve=1.2406940, (0 missing)
## bucket2008 < 3.5 to the left, improve=0.2455781, (0 missing)
## Surrogate splits:
## reimbursement2008 < 14660 to the right, agree=0.679, adj=0.054, (0 split)
##
## Node number 247699: 60 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0002183382
## class counts: 16 30 8 6 0
## probabilities: 0.267 0.500 0.133 0.100 0.000
##
## Node number 254796: 112 observations
## predicted class=B1 expected loss=0.6160714 P(node) =0.0004075647
## class counts: 43 27 31 10 1
## probabilities: 0.384 0.241 0.277 0.089 0.009
##
## Node number 254797: 89 observations
## predicted class=B2 expected loss=0.6516854 P(node) =0.0003238684
## class counts: 20 31 18 17 3
## probabilities: 0.225 0.348 0.202 0.191 0.034
##
## Node number 494276: 135 observations
## predicted class=B1 expected loss=0.5259259 P(node) =0.000491261
## class counts: 64 49 20 2 0
## probabilities: 0.474 0.363 0.148 0.015 0.000
##
## Node number 494277: 457 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5820569 P(node) =0.00166301
## class counts: 184 191 59 23 0
## probabilities: 0.403 0.418 0.129 0.050 0.000
## left son=988554 (290 obs) right son=988555 (167 obs)
## Primary splits:
## age < 74.5 to the left, improve=0.9874503, (0 missing)
## reimbursement2008 < 3495 to the left, improve=0.9861916, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3272674, (0 missing)
## stroke < 0.5 to the left, improve=0.2337493, (0 missing)
## copd < 0.5 to the left, improve=0.1994562, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.637, adj=0.006, (0 split)
##
## Node number 494284: 114 observations
## predicted class=B1 expected loss=0.4824561 P(node) =0.0004148426
## class counts: 59 32 18 4 1
## probabilities: 0.518 0.281 0.158 0.035 0.009
##
## Node number 494285: 10 observations
## predicted class=B2 expected loss=0.3 P(node) =3.63897e-05
## class counts: 1 7 1 1 0
## probabilities: 0.100 0.700 0.100 0.100 0.000
##
## Node number 494286: 217 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.0007896566
## class counts: 78 93 30 15 1
## probabilities: 0.359 0.429 0.138 0.069 0.005
##
## Node number 494287: 32 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.625 P(node) =0.0001164471
## class counts: 12 7 12 1 0
## probabilities: 0.375 0.219 0.375 0.031 0.000
## left son=988574 (11 obs) right son=988575 (21 obs)
## Primary splits:
## age < 72.5 to the left, improve=1.8097940, (0 missing)
## reimbursement2008 < 5975 to the left, improve=0.7232143, (0 missing)
## copd < 0.5 to the left, improve=0.6875000, (0 missing)
##
## Node number 495396: 72 observations
## predicted class=B1 expected loss=0.5138889 P(node) =0.0002620059
## class counts: 35 17 12 8 0
## probabilities: 0.486 0.236 0.167 0.111 0.000
##
## Node number 495397: 37 observations
## predicted class=B2 expected loss=0.5405405 P(node) =0.0001346419
## class counts: 11 17 7 2 0
## probabilities: 0.297 0.459 0.189 0.054 0.000
##
## Node number 988554: 290 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5724138 P(node) =0.001055301
## class counts: 124 114 37 15 0
## probabilities: 0.428 0.393 0.128 0.052 0.000
## left son=1977108 (234 obs) right son=1977109 (56 obs)
## Primary splits:
## age < 62.5 to the right, improve=1.0825800, (0 missing)
## reimbursement2008 < 3945 to the right, improve=0.7040408, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.6026089, (0 missing)
## stroke < 0.5 to the left, improve=0.2655768, (0 missing)
## copd < 0.5 to the left, improve=0.1804923, (0 missing)
##
## Node number 988555: 167 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5389222 P(node) =0.0006077081
## class counts: 60 77 22 8 0
## probabilities: 0.359 0.461 0.132 0.048 0.000
## left son=1977110 (39 obs) right son=1977111 (128 obs)
## Primary splits:
## reimbursement2008 < 4105 to the right, improve=1.3886510, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7439669, (0 missing)
## age < 81.5 to the left, improve=0.4824922, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2060442, (0 missing)
## copd < 0.5 to the left, improve=0.1297289, (0 missing)
##
## Node number 988574: 11 observations
## predicted class=B1 expected loss=0.3636364 P(node) =4.002868e-05
## class counts: 7 2 2 0 0
## probabilities: 0.636 0.182 0.182 0.000 0.000
##
## Node number 988575: 21 observations
## predicted class=B3 expected loss=0.5238095 P(node) =7.641838e-05
## class counts: 5 5 10 1 0
## probabilities: 0.238 0.238 0.476 0.048 0.000
##
## Node number 1977108: 234 observations
## predicted class=B1 expected loss=0.5470085 P(node) =0.0008515191
## class counts: 106 89 28 11 0
## probabilities: 0.453 0.380 0.120 0.047 0.000
##
## Node number 1977109: 56 observations
## predicted class=B2 expected loss=0.5535714 P(node) =0.0002037823
## class counts: 18 25 9 4 0
## probabilities: 0.321 0.446 0.161 0.071 0.000
##
## Node number 1977110: 39 observations
## predicted class=B1 expected loss=0.5128205 P(node) =0.0001419198
## class counts: 19 14 5 1 0
## probabilities: 0.487 0.359 0.128 0.026 0.000
##
## Node number 1977111: 128 observations
## predicted class=B2 expected loss=0.5078125 P(node) =0.0004657882
## class counts: 41 63 17 7 0
## probabilities: 0.320 0.492 0.133 0.055 0.000
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) reimbursement2008< 1565 165987 20938 B1 (0.87 0.074 0.037 0.014 0.0014) *
## 3) reimbursement2008>=1565 108816 68841 B2 (0.36 0.37 0.17 0.088 0.012)
## 6) reimbursement2008< 3065 39298 18853 B1 (0.52 0.31 0.12 0.045 0.0046)
## 12) reimbursement2008< 2175 20077 8527 B1 (0.58 0.27 0.11 0.042 0.0038)
## 24) diabetes< 0.5 8826 3280 B1 (0.63 0.24 0.091 0.035 0.0029) *
## 25) diabetes>=0.5 11251 5247 B1 (0.53 0.29 0.12 0.046 0.0045)
## 50) kidney< 0.5 9007 4045 B1 (0.55 0.28 0.12 0.042 0.0044)
## 100) reimbursement2008< 1875 4935 2071 B1 (0.58 0.26 0.11 0.042 0.0045) *
## 101) reimbursement2008>=1875 4072 1974 B1 (0.52 0.31 0.13 0.042 0.0044)
## 202) cancer< 0.5 3786 1807 B1 (0.52 0.3 0.13 0.041 0.0048) *
## 203) cancer>=0.5 286 167 B1 (0.42 0.4 0.13 0.056 0)
## 406) age< 73.5 128 64 B1 (0.5 0.33 0.11 0.062 0)
## 812) depression< 0.5 95 41 B1 (0.57 0.26 0.12 0.053 0) *
## 813) depression>=0.5 33 16 B2 (0.3 0.52 0.091 0.091 0) *
## 407) age>=73.5 158 85 B2 (0.35 0.46 0.14 0.051 0) *
## 51) kidney>=0.5 2244 1202 B1 (0.46 0.33 0.14 0.064 0.0049)
## 102) heart.failure< 0.5 992 477 B1 (0.52 0.29 0.13 0.057 0.002) *
## 103) heart.failure>=0.5 1252 725 B1 (0.42 0.36 0.15 0.069 0.0072)
## 206) arthritis< 0.5 904 486 B1 (0.46 0.34 0.13 0.063 0.0066)
## 412) reimbursement2008< 1735 270 125 B1 (0.54 0.27 0.13 0.056 0.0037) *
## 413) reimbursement2008>=1735 634 361 B1 (0.43 0.36 0.13 0.066 0.0079)
## 826) age< 91.5 596 330 B1 (0.45 0.36 0.13 0.057 0.0084)
## 1652) reimbursement2008>=1765 555 302 B1 (0.46 0.35 0.14 0.052 0.0072)
## 3304) reimbursement2008< 1955 265 130 B1 (0.51 0.31 0.13 0.049 0.0038)
## 6608) stroke< 0.5 251 118 B1 (0.53 0.29 0.13 0.044 0.004)
## 13216) cancer< 0.5 235 106 B1 (0.55 0.27 0.13 0.047 0.0043) *
## 13217) cancer>=0.5 16 7 B2 (0.25 0.56 0.19 0 0) *
## 6609) stroke>=0.5 14 5 B2 (0.14 0.64 0.071 0.14 0) *
## 3305) reimbursement2008>=1955 290 172 B1 (0.41 0.38 0.14 0.055 0.01)
## 6610) age< 81.5 213 120 B1 (0.44 0.35 0.14 0.07 0.0047)
## 13220) age>=44.5 201 110 B1 (0.45 0.33 0.14 0.07 0) *
## 13221) age< 44.5 12 5 B2 (0.17 0.58 0.083 0.083 0.083) *
## 6611) age>=81.5 77 40 B2 (0.32 0.48 0.16 0.013 0.026) *
## 1653) reimbursement2008< 1765 41 20 B2 (0.32 0.51 0.024 0.12 0.024) *
## 827) age>=91.5 38 21 B2 (0.18 0.45 0.16 0.21 0) *
## 207) arthritis>=0.5 348 205 B2 (0.31 0.41 0.18 0.086 0.0086) *
## 13) reimbursement2008>=2175 19221 10326 B1 (0.46 0.35 0.13 0.049 0.0054)
## 26) diabetes< 0.5 7137 3360 B1 (0.53 0.31 0.11 0.042 0.0046)
## 52) arthritis< 0.5 5554 2471 B1 (0.56 0.3 0.1 0.039 0.0049)
## 104) ihd< 0.5 2348 933 B1 (0.6 0.27 0.092 0.031 0.0051) *
## 105) ihd>=0.5 3206 1538 B1 (0.52 0.32 0.11 0.045 0.0047)
## 210) depression< 0.5 2325 1056 B1 (0.55 0.3 0.11 0.043 0.0052) *
## 211) depression>=0.5 881 482 B1 (0.45 0.36 0.13 0.052 0.0034)
## 422) kidney< 0.5 763 405 B1 (0.47 0.34 0.13 0.052 0.0039) *
## 423) kidney>=0.5 118 63 B2 (0.35 0.47 0.14 0.051 0)
## 846) reimbursement2008>=2865 22 10 B1 (0.55 0.23 0.18 0.045 0) *
## 847) reimbursement2008< 2865 96 46 B2 (0.3 0.52 0.12 0.052 0) *
## 53) arthritis>=0.5 1583 889 B1 (0.44 0.37 0.14 0.052 0.0038)
## 106) stroke< 0.5 1525 844 B1 (0.45 0.36 0.13 0.054 0.0039)
## 212) cancer< 0.5 1438 784 B1 (0.45 0.36 0.13 0.053 0.0042)
## 424) reimbursement2008>=2715 495 280 B1 (0.43 0.41 0.1 0.053 0.004)
## 848) reimbursement2008>=2795 385 210 B1 (0.45 0.38 0.1 0.06 0.0052)
## 1696) age< 80.5 263 131 B1 (0.5 0.36 0.099 0.042 0) *
## 1697) age>=80.5 122 70 B2 (0.35 0.43 0.11 0.098 0.016) *
## 849) reimbursement2008< 2795 110 54 B2 (0.36 0.51 0.1 0.027 0) *
## 425) reimbursement2008< 2715 943 504 B1 (0.47 0.33 0.15 0.053 0.0042) *
## 213) cancer>=0.5 87 48 B2 (0.31 0.45 0.17 0.069 0) *
## 107) stroke>=0.5 58 26 B2 (0.22 0.55 0.21 0.017 0) *
## 27) diabetes>=0.5 12084 6966 B1 (0.42 0.37 0.15 0.054 0.0059)
## 54) arthritis< 0.5 8413 4653 B1 (0.45 0.35 0.15 0.052 0.0056)
## 108) heart.failure< 0.5 4375 2220 B1 (0.49 0.34 0.13 0.039 0.0039)
## 216) cancer< 0.5 3992 1978 B1 (0.5 0.33 0.12 0.038 0.0033)
## 432) ihd< 0.5 1265 562 B1 (0.56 0.29 0.12 0.035 0.0032) *
## 433) ihd>=0.5 2727 1416 B1 (0.48 0.35 0.13 0.04 0.0033)
## 866) reimbursement2008< 2615 1499 736 B1 (0.51 0.33 0.13 0.035 0.002) *
## 867) reimbursement2008>=2615 1228 680 B1 (0.45 0.37 0.13 0.046 0.0049)
## 1734) reimbursement2008>=2995 171 81 B1 (0.53 0.27 0.14 0.058 0.0058) *
## 1735) reimbursement2008< 2995 1057 599 B1 (0.43 0.39 0.13 0.044 0.0047)
## 3470) age< 83.5 840 458 B1 (0.45 0.37 0.12 0.049 0.006)
## 6940) age< 54.5 71 31 B1 (0.56 0.23 0.15 0.042 0.014) *
## 6941) age>=54.5 769 427 B1 (0.44 0.38 0.12 0.049 0.0052)
## 13882) age>=70.5 472 249 B1 (0.47 0.35 0.12 0.053 0.0085)
## 27764) age>=73.5 343 191 B1 (0.44 0.4 0.11 0.047 0.0058)
## 55528) reimbursement2008>=2835 117 57 B1 (0.51 0.32 0.13 0.026 0.0085)
## 111056) reimbursement2008< 2945 78 32 B1 (0.59 0.24 0.14 0.026 0) *
## 111057) reimbursement2008>=2945 39 20 B2 (0.36 0.49 0.1 0.026 0.026) *
## 55529) reimbursement2008< 2835 226 128 B2 (0.41 0.43 0.097 0.058 0.0044)
## 111058) age>=80.5 72 35 B1 (0.51 0.4 0.069 0.014 0) *
## 111059) age< 80.5 154 85 B2 (0.36 0.45 0.11 0.078 0.0065) *
## 27765) age< 73.5 129 58 B1 (0.55 0.21 0.16 0.07 0.016) *
## 13883) age< 70.5 297 167 B2 (0.4 0.44 0.12 0.044 0)
## 27766) osteoporosis< 0.5 218 125 B1 (0.43 0.41 0.13 0.037 0)
## 55532) reimbursement2008< 2945 194 108 B1 (0.44 0.38 0.14 0.036 0) *
## 55533) reimbursement2008>=2945 24 9 B2 (0.29 0.62 0.042 0.042 0) *
## 27767) osteoporosis>=0.5 79 38 B2 (0.33 0.52 0.089 0.063 0) *
## 3471) age>=83.5 217 116 B2 (0.35 0.47 0.16 0.028 0) *
## 217) cancer>=0.5 383 220 B2 (0.37 0.43 0.15 0.044 0.01)
## 434) reimbursement2008< 2705 238 136 B1 (0.43 0.36 0.16 0.038 0.013)
## 868) depression< 0.5 167 84 B1 (0.5 0.3 0.15 0.042 0.012) *
## 869) depression>=0.5 71 35 B2 (0.27 0.51 0.18 0.028 0.014) *
## 435) reimbursement2008>=2705 145 68 B2 (0.27 0.53 0.14 0.055 0.0069) *
## 109) heart.failure>=0.5 4038 2433 B1 (0.4 0.36 0.17 0.066 0.0074)
## 218) kidney< 0.5 2819 1620 B1 (0.43 0.35 0.16 0.065 0.0064)
## 436) ihd< 0.5 635 319 B1 (0.5 0.31 0.15 0.041 0.0063) *
## 437) ihd>=0.5 2184 1301 B1 (0.4 0.36 0.16 0.072 0.0064)
## 874) reimbursement2008< 2315 393 202 B1 (0.49 0.34 0.12 0.051 0.0051) *
## 875) reimbursement2008>=2315 1791 1099 B1 (0.39 0.36 0.17 0.076 0.0067)
## 1750) age>=39.5 1752 1066 B1 (0.39 0.36 0.17 0.075 0.0068)
## 3500) depression< 0.5 1099 639 B1 (0.42 0.35 0.15 0.069 0.0064)
## 7000) age< 95.5 1074 620 B1 (0.42 0.35 0.15 0.069 0.0065)
## 14000) copd< 0.5 808 450 B1 (0.44 0.34 0.14 0.075 0.0062) *
## 14001) copd>=0.5 266 166 B2 (0.36 0.38 0.21 0.049 0.0075)
## 28002) reimbursement2008>=2540 192 114 B2 (0.38 0.41 0.15 0.057 0.01)
## 56004) age< 78.5 124 67 B1 (0.46 0.38 0.13 0.032 0)
## 112008) age>=72.5 46 20 B1 (0.57 0.22 0.17 0.043 0) *
## 112009) age< 72.5 78 41 B2 (0.4 0.47 0.1 0.026 0) *
## 56005) age>=78.5 68 37 B2 (0.22 0.46 0.19 0.1 0.029) *
## 28003) reimbursement2008< 2540 74 48 B3 (0.32 0.3 0.35 0.027 0)
## 56006) cancer>=0.5 8 0 B2 (0 1 0 0 0) *
## 56007) cancer< 0.5 66 40 B3 (0.36 0.21 0.39 0.03 0)
## 112014) alzheimers< 0.5 40 24 B1 (0.4 0.3 0.27 0.025 0) *
## 112015) alzheimers>=0.5 26 11 B3 (0.31 0.077 0.58 0.038 0) *
## 7001) age>=95.5 25 10 B2 (0.24 0.6 0.08 0.08 0) *
## 3501) depression>=0.5 653 412 B2 (0.35 0.37 0.19 0.084 0.0077)
## 7002) reimbursement2008< 2655 303 183 B1 (0.4 0.33 0.2 0.069 0.0033) *
## 7003) reimbursement2008>=2655 350 208 B2 (0.3 0.41 0.18 0.097 0.011) *
## 1751) age< 39.5 39 18 B2 (0.15 0.54 0.15 0.15 0) *
## 219) kidney>=0.5 1219 734 B2 (0.33 0.4 0.19 0.07 0.0098)
## 438) reimbursement2008< 2615 613 379 B1 (0.38 0.37 0.18 0.059 0.0098)
## 876) osteoporosis>=0.5 180 102 B1 (0.43 0.38 0.13 0.061 0)
## 1752) reimbursement2008< 2455 112 56 B1 (0.5 0.29 0.12 0.08 0) *
## 1753) reimbursement2008>=2455 68 33 B2 (0.32 0.51 0.13 0.029 0) *
## 877) osteoporosis< 0.5 433 275 B2 (0.36 0.36 0.2 0.058 0.014)
## 1754) stroke< 0.5 403 252 B1 (0.37 0.36 0.2 0.05 0.015)
## 3508) reimbursement2008< 2585 382 238 B2 (0.37 0.38 0.19 0.047 0.01)
## 7016) depression< 0.5 229 136 B1 (0.41 0.36 0.19 0.035 0.0087)
## 14032) cancer>=0.5 15 5 B1 (0.67 0.13 0.2 0 0) *
## 14033) cancer< 0.5 214 131 B1 (0.39 0.37 0.19 0.037 0.0093)
## 28066) reimbursement2008< 2515 169 98 B1 (0.42 0.34 0.19 0.036 0.012) *
## 28067) reimbursement2008>=2515 45 23 B2 (0.27 0.49 0.2 0.044 0) *
## 7017) depression>=0.5 153 91 B2 (0.32 0.41 0.2 0.065 0.013)
## 14034) reimbursement2008>=2545 14 5 B1 (0.64 0.14 0.14 0.071 0) *
## 14035) reimbursement2008< 2545 139 79 B2 (0.29 0.43 0.2 0.065 0.014) *
## 3509) reimbursement2008>=2585 21 12 B1 (0.43 0.14 0.24 0.095 0.095) *
## 1755) stroke>=0.5 30 19 B2 (0.17 0.37 0.3 0.17 0) *
## 439) reimbursement2008>=2615 606 347 B2 (0.28 0.43 0.2 0.081 0.0099) *
## 55) arthritis>=0.5 3671 2129 B2 (0.37 0.42 0.15 0.057 0.0065)
## 110) reimbursement2008< 2665 2068 1224 B1 (0.41 0.4 0.14 0.057 0.0048)
## 220) ihd< 0.5 517 274 B1 (0.47 0.37 0.11 0.048 0.0019)
## 440) reimbursement2008< 2295 143 57 B1 (0.6 0.26 0.077 0.063 0) *
## 441) reimbursement2008>=2295 374 217 B1 (0.42 0.41 0.12 0.043 0.0027)
## 882) reimbursement2008< 2315 25 6 B2 (0.24 0.76 0 0 0) *
## 883) reimbursement2008>=2315 349 198 B1 (0.43 0.39 0.13 0.046 0.0029)
## 1766) cancer< 0.5 336 186 B1 (0.45 0.38 0.13 0.042 0)
## 3532) age< 90.5 322 176 B1 (0.45 0.37 0.13 0.043 0) *
## 3533) age>=90.5 14 5 B2 (0.29 0.64 0.071 0 0) *
## 1767) cancer>=0.5 13 6 B2 (0.077 0.54 0.15 0.15 0.077) *
## 221) ihd>=0.5 1551 925 B2 (0.39 0.4 0.14 0.059 0.0058)
## 442) age< 35 18 5 B1 (0.72 0.22 0 0.056 0) *
## 443) age>=35 1533 911 B2 (0.38 0.41 0.15 0.059 0.0059)
## 886) kidney< 0.5 1101 656 B1 (0.4 0.4 0.14 0.052 0.0045)
## 1772) stroke< 0.5 1057 623 B1 (0.41 0.4 0.14 0.051 0.0038)
## 3544) cancer< 0.5 1008 590 B1 (0.41 0.4 0.13 0.053 0.004)
## 7088) reimbursement2008>=2535 275 154 B2 (0.39 0.44 0.12 0.04 0.0036)
## 14176) age< 63.5 44 17 B2 (0.32 0.61 0.045 0.023 0) *
## 14177) age>=63.5 231 137 B1 (0.41 0.41 0.14 0.043 0.0043)
## 28354) alzheimers< 0.5 169 95 B1 (0.44 0.36 0.15 0.047 0.0059) *
## 28355) alzheimers>=0.5 62 28 B2 (0.32 0.55 0.097 0.032 0) *
## 7089) reimbursement2008< 2535 733 423 B1 (0.42 0.38 0.14 0.057 0.0041)
## 14178) age>=97.5 10 3 B1 (0.7 0.2 0.1 0 0) *
## 14179) age< 97.5 723 420 B1 (0.42 0.38 0.14 0.058 0.0041)
## 28358) age< 90.5 689 394 B1 (0.43 0.38 0.13 0.055 0.0044)
## 56716) heart.failure< 0.5 367 198 B1 (0.46 0.36 0.14 0.035 0.0082) *
## 56717) heart.failure>=0.5 322 192 B2 (0.39 0.4 0.13 0.078 0)
## 113434) age< 67.5 78 43 B1 (0.45 0.29 0.19 0.064 0) *
## 113435) age>=67.5 244 137 B2 (0.37 0.44 0.11 0.082 0) *
## 28359) age>=90.5 34 18 B2 (0.24 0.47 0.18 0.12 0) *
## 3545) cancer>=0.5 49 29 B2 (0.33 0.41 0.24 0.02 0) *
## 1773) stroke>=0.5 44 20 B2 (0.25 0.55 0.11 0.068 0.023) *
## 887) kidney>=0.5 432 254 B2 (0.33 0.41 0.17 0.079 0.0093)
## 1774) reimbursement2008>=2215 403 232 B2 (0.32 0.42 0.17 0.069 0.0099) *
## 1775) reimbursement2008< 2215 29 16 B1 (0.45 0.24 0.1 0.21 0) *
## 111) reimbursement2008>=2665 1603 878 B2 (0.32 0.45 0.16 0.058 0.0087) *
## 7) reimbursement2008>=3065 69518 41677 B2 (0.27 0.4 0.2 0.11 0.017)
## 14) diabetes< 0.5 15717 8966 B1 (0.43 0.35 0.15 0.064 0.0071)
## 28) cancer< 0.5 13123 7034 B1 (0.46 0.34 0.13 0.058 0.0065)
## 56) arthritis< 0.5 9625 4692 B1 (0.51 0.31 0.12 0.054 0.0058)
## 112) ihd< 0.5 3135 1246 B1 (0.6 0.26 0.095 0.036 0.0032)
## 224) depression< 0.5 2292 821 B1 (0.64 0.24 0.08 0.034 0.0044) *
## 225) depression>=0.5 843 425 B1 (0.5 0.33 0.14 0.04 0)
## 450) age< 92.5 810 398 B1 (0.51 0.32 0.13 0.041 0)
## 900) reimbursement2008>=11525 117 40 B1 (0.66 0.21 0.068 0.06 0) *
## 901) reimbursement2008< 11525 693 358 B1 (0.48 0.33 0.14 0.038 0)
## 1802) reimbursement2008< 11105 684 352 B1 (0.49 0.34 0.14 0.038 0)
## 3604) reimbursement2008< 4365 286 134 B1 (0.53 0.33 0.12 0.017 0) *
## 3605) reimbursement2008>=4365 398 218 B1 (0.45 0.34 0.15 0.053 0)
## 7210) reimbursement2008>=4700 340 173 B1 (0.49 0.31 0.14 0.053 0) *
## 7211) reimbursement2008< 4700 58 28 B2 (0.22 0.52 0.21 0.052 0) *
## 1803) reimbursement2008>=11105 9 4 B3 (0.33 0.11 0.56 0 0) *
## 451) age>=92.5 33 14 B2 (0.18 0.58 0.21 0.03 0) *
## 113) ihd>=0.5 6490 3446 B1 (0.47 0.33 0.13 0.063 0.0071)
## 226) depression< 0.5 4266 2110 B1 (0.51 0.31 0.12 0.056 0.0061)
## 452) osteoporosis< 0.5 3304 1572 B1 (0.52 0.3 0.12 0.055 0.0064)
## 904) reimbursement2008>=5905 1626 714 B1 (0.56 0.25 0.12 0.061 0.0068) *
## 905) reimbursement2008< 5905 1678 858 B1 (0.49 0.34 0.12 0.05 0.006)
## 1810) reimbursement2008< 5695 1608 814 B1 (0.49 0.33 0.12 0.051 0.0062) *
## 1811) reimbursement2008>=5695 70 34 B2 (0.37 0.51 0.086 0.029 0) *
## 453) osteoporosis>=0.5 962 538 B1 (0.44 0.38 0.12 0.057 0.0052)
## 906) stroke< 0.5 857 465 B1 (0.46 0.37 0.12 0.056 0.0047)
## 1812) heart.failure< 0.5 405 203 B1 (0.5 0.35 0.1 0.044 0.0074)
## 3624) age< 83.5 329 159 B1 (0.52 0.32 0.11 0.049 0.0091) *
## 3625) age>=83.5 76 41 B2 (0.42 0.46 0.092 0.026 0)
## 7250) reimbursement2008>=6785 21 7 B1 (0.67 0.24 0.095 0 0) *
## 7251) reimbursement2008< 6785 55 25 B2 (0.33 0.55 0.091 0.036 0) *
## 1813) heart.failure>=0.5 452 262 B1 (0.42 0.38 0.13 0.066 0.0022)
## 3626) reimbursement2008>=3875 362 201 B1 (0.44 0.35 0.13 0.069 0.0028) *
## 3627) reimbursement2008< 3875 90 45 B2 (0.32 0.5 0.12 0.056 0)
## 7254) age< 69.5 21 9 B1 (0.57 0.29 0.048 0.095 0) *
## 7255) age>=69.5 69 30 B2 (0.25 0.57 0.14 0.043 0) *
## 907) stroke>=0.5 105 54 B2 (0.3 0.49 0.13 0.067 0.0095) *
## 227) depression>=0.5 2224 1336 B1 (0.4 0.35 0.16 0.076 0.009)
## 454) kidney< 0.5 1518 863 B1 (0.43 0.34 0.16 0.061 0.0053) *
## 455) kidney>=0.5 706 440 B2 (0.33 0.38 0.17 0.11 0.017)
## 910) reimbursement2008>=3155 696 431 B2 (0.33 0.38 0.16 0.11 0.017)
## 1820) heart.failure< 0.5 177 99 B1 (0.44 0.35 0.15 0.062 0) *
## 1821) heart.failure>=0.5 519 316 B2 (0.3 0.39 0.17 0.12 0.023) *
## 911) reimbursement2008< 3155 10 4 B3 (0.1 0.1 0.6 0.2 0) *
## 57) arthritis>=0.5 3498 2017 B2 (0.33 0.42 0.17 0.069 0.0083)
## 114) reimbursement2008< 8525 2340 1270 B2 (0.31 0.46 0.17 0.062 0.0064)
## 228) reimbursement2008< 4645 1359 754 B2 (0.34 0.45 0.15 0.056 0.0059)
## 456) ihd< 0.5 440 248 B2 (0.4 0.44 0.11 0.045 0.0045)
## 912) reimbursement2008< 3155 58 22 B2 (0.34 0.62 0 0.017 0.017) *
## 913) reimbursement2008>=3155 382 225 B1 (0.41 0.41 0.13 0.05 0.0026)
## 1826) reimbursement2008< 3245 25 8 B1 (0.68 0.28 0.04 0 0) *
## 1827) reimbursement2008>=3245 357 208 B2 (0.39 0.42 0.13 0.053 0.0028)
## 3654) age>=80.5 91 46 B1 (0.49 0.34 0.099 0.066 0) *
## 3655) age< 80.5 266 148 B2 (0.36 0.44 0.15 0.049 0.0038) *
## 457) ihd>=0.5 919 506 B2 (0.32 0.45 0.17 0.061 0.0065) *
## 229) reimbursement2008>=4645 981 516 B2 (0.26 0.47 0.19 0.069 0.0071) *
## 115) reimbursement2008>=8525 1158 722 B1 (0.38 0.35 0.17 0.085 0.012)
## 230) copd< 0.5 714 396 B1 (0.45 0.33 0.13 0.085 0.007)
## 460) depression< 0.5 412 196 B1 (0.52 0.29 0.1 0.073 0.0097) *
## 461) depression>=0.5 302 183 B2 (0.34 0.39 0.16 0.1 0.0033)
## 922) age>=92.5 9 3 B1 (0.67 0 0.11 0.11 0.11) *
## 923) age< 92.5 293 174 B2 (0.33 0.41 0.16 0.1 0)
## 1846) stroke>=0.5 39 19 B1 (0.51 0.31 0.1 0.077 0) *
## 1847) stroke< 0.5 254 147 B2 (0.3 0.42 0.17 0.11 0) *
## 231) copd>=0.5 444 272 B2 (0.27 0.39 0.24 0.086 0.02)
## 462) osteoporosis< 0.5 282 187 B2 (0.31 0.34 0.25 0.082 0.018)
## 924) reimbursement2008< 27390 220 143 B1 (0.35 0.3 0.26 0.073 0.018)
## 1848) reimbursement2008>=12810 132 78 B1 (0.41 0.32 0.2 0.068 0.0076)
## 3696) age< 84.5 105 55 B1 (0.48 0.33 0.15 0.029 0.0095) *
## 3697) age>=84.5 27 17 B3 (0.15 0.26 0.37 0.22 0) *
## 1849) reimbursement2008< 12810 88 57 B3 (0.26 0.27 0.35 0.08 0.034) *
## 925) reimbursement2008>=27390 62 33 B2 (0.18 0.47 0.23 0.11 0.016) *
## 463) osteoporosis>=0.5 162 85 B2 (0.19 0.48 0.22 0.093 0.025) *
## 29) cancer>=0.5 2594 1539 B2 (0.26 0.41 0.24 0.091 0.01)
## 58) reimbursement2008< 5770 1000 562 B2 (0.3 0.44 0.19 0.07 0.005) *
## 59) reimbursement2008>=5770 1594 977 B2 (0.23 0.39 0.27 0.1 0.014)
## 118) reimbursement2008>=8645 1054 656 B2 (0.27 0.38 0.24 0.1 0.015)
## 236) arthritis< 0.5 745 464 B2 (0.31 0.38 0.2 0.097 0.013)
## 472) ihd< 0.5 159 94 B1 (0.41 0.32 0.21 0.05 0.013)
## 944) reimbursement2008>=11995 76 36 B1 (0.53 0.24 0.16 0.066 0.013) *
## 945) reimbursement2008< 11995 83 50 B2 (0.3 0.4 0.25 0.036 0.012) *
## 473) ihd>=0.5 586 356 B2 (0.28 0.39 0.2 0.11 0.014) *
## 237) arthritis>=0.5 309 192 B2 (0.16 0.38 0.32 0.12 0.019)
## 474) reimbursement2008>=10960 237 136 B2 (0.15 0.43 0.3 0.11 0.013)
## 948) copd< 0.5 126 64 B2 (0.17 0.49 0.26 0.071 0) *
## 949) copd>=0.5 111 72 B2 (0.12 0.35 0.35 0.15 0.027)
## 1898) age< 75.5 54 30 B3 (0.15 0.26 0.44 0.13 0.019) *
## 1899) age>=75.5 57 32 B2 (0.088 0.44 0.26 0.18 0.035) *
## 475) reimbursement2008< 10960 72 44 B3 (0.19 0.22 0.39 0.15 0.042) *
## 119) reimbursement2008< 8645 540 321 B2 (0.16 0.41 0.32 0.11 0.011)
## 238) heart.failure>=0.5 243 128 B2 (0.14 0.47 0.28 0.099 0.016) *
## 239) heart.failure< 0.5 297 191 B3 (0.18 0.35 0.36 0.11 0.0067)
## 478) depression< 0.5 226 141 B2 (0.18 0.38 0.33 0.12 0.0044) *
## 479) depression>=0.5 71 39 B3 (0.17 0.27 0.45 0.099 0.014) *
## 15) diabetes>=0.5 53801 31450 B2 (0.23 0.42 0.21 0.13 0.02)
## 30) kidney< 0.5 25067 14311 B2 (0.3 0.43 0.19 0.076 0.0074)
## 60) arthritis< 0.5 15178 9179 B2 (0.35 0.4 0.17 0.069 0.0063)
## 120) cancer< 0.5 12572 7709 B2 (0.39 0.39 0.16 0.063 0.0059)
## 240) ihd< 0.5 2617 1376 B1 (0.47 0.34 0.13 0.049 0.0053)
## 480) reimbursement2008>=9400 403 171 B1 (0.58 0.21 0.15 0.05 0.0099) *
## 481) reimbursement2008< 9400 2214 1205 B1 (0.46 0.36 0.13 0.048 0.0045)
## 962) osteoporosis< 0.5 1636 847 B1 (0.48 0.34 0.12 0.049 0.0043)
## 1924) alzheimers< 0.5 1127 559 B1 (0.5 0.33 0.12 0.042 0.0035) *
## 1925) alzheimers>=0.5 509 288 B1 (0.43 0.36 0.13 0.065 0.0059)
## 3850) reimbursement2008< 3775 137 68 B1 (0.5 0.3 0.12 0.066 0.0073) *
## 3851) reimbursement2008>=3775 372 220 B1 (0.41 0.39 0.13 0.065 0.0054)
## 7702) reimbursement2008>=4055 330 188 B1 (0.43 0.36 0.13 0.07 0.0061)
## 15404) reimbursement2008>=4185 309 177 B1 (0.43 0.38 0.12 0.065 0.0065)
## 30808) reimbursement2008>=4635 253 138 B1 (0.45 0.35 0.12 0.067 0.0079)
## 61616) age< 96 245 131 B1 (0.47 0.36 0.11 0.065 0.0041)
## 123232) reimbursement2008< 8170 209 106 B1 (0.49 0.33 0.11 0.067 0) *
## 123233) reimbursement2008>=8170 36 19 B2 (0.31 0.47 0.14 0.056 0.028)
## 246466) age< 74.5 15 6 B1 (0.6 0.13 0.2 0 0.067) *
## 246467) age>=74.5 21 6 B2 (0.095 0.71 0.095 0.095 0) *
## 61617) age>=96 8 5 B3 (0.12 0.25 0.38 0.12 0.12) *
## 30809) reimbursement2008< 4635 56 28 B2 (0.3 0.5 0.14 0.054 0) *
## 15405) reimbursement2008< 4185 21 11 B1 (0.48 0.095 0.29 0.14 0) *
## 7703) reimbursement2008< 4055 42 17 B2 (0.24 0.6 0.14 0.024 0) *
## 963) osteoporosis>=0.5 578 342 B2 (0.38 0.41 0.16 0.047 0.0052)
## 1926) depression< 0.5 339 189 B1 (0.44 0.37 0.13 0.047 0.0029)
## 3852) reimbursement2008< 4905 211 119 B1 (0.44 0.42 0.11 0.033 0)
## 7704) reimbursement2008< 4075 142 71 B1 (0.5 0.36 0.11 0.035 0) *
## 7705) reimbursement2008>=4075 69 31 B2 (0.3 0.55 0.12 0.029 0) *
## 3853) reimbursement2008>=4905 128 70 B1 (0.45 0.3 0.17 0.07 0.0078) *
## 1927) depression>=0.5 239 130 B2 (0.29 0.46 0.2 0.046 0.0084)
## 3854) copd< 0.5 181 88 B2 (0.31 0.51 0.13 0.039 0.011) *
## 3855) copd>=0.5 58 34 B3 (0.24 0.28 0.41 0.069 0) *
## 241) ihd>=0.5 9955 5976 B2 (0.36 0.4 0.17 0.067 0.006)
## 482) depression< 0.5 5563 3339 B1 (0.4 0.38 0.15 0.059 0.0059)
## 964) reimbursement2008>=8955 1363 758 B1 (0.44 0.32 0.16 0.067 0.0088)
## 1928) copd< 0.5 798 405 B1 (0.49 0.32 0.12 0.064 0.0038) *
## 1929) copd>=0.5 565 353 B1 (0.38 0.32 0.22 0.073 0.016)
## 3858) stroke>=0.5 116 64 B2 (0.35 0.45 0.12 0.06 0.017)
## 7716) age>=74.5 63 36 B1 (0.43 0.33 0.16 0.063 0.016) *
## 7717) age< 74.5 53 22 B2 (0.26 0.58 0.075 0.057 0.019) *
## 3859) stroke< 0.5 449 278 B1 (0.38 0.28 0.24 0.076 0.016) *
## 965) reimbursement2008< 8955 4200 2510 B2 (0.39 0.4 0.15 0.056 0.005)
## 1930) heart.failure< 0.5 1953 1129 B1 (0.42 0.4 0.13 0.045 0.0041)
## 3860) reimbursement2008< 3415 343 172 B1 (0.5 0.37 0.096 0.032 0.0058) *
## 3861) reimbursement2008>=3415 1610 954 B2 (0.41 0.41 0.14 0.048 0.0037)
## 7722) age< 42.5 43 17 B1 (0.6 0.26 0.07 0.07 0) *
## 7723) age>=42.5 1567 922 B2 (0.4 0.41 0.14 0.047 0.0038)
## 15446) age>=50.5 1527 894 B2 (0.4 0.41 0.13 0.047 0.0039)
## 30892) reimbursement2008>=3465 1478 870 B2 (0.41 0.41 0.13 0.045 0.0027)
## 61784) reimbursement2008< 4655 759 431 B1 (0.43 0.4 0.12 0.043 0.0013)
## 123568) reimbursement2008>=4315 158 79 B1 (0.5 0.35 0.095 0.051 0.0063) *
## 123569) reimbursement2008< 4315 601 352 B1 (0.41 0.41 0.13 0.042 0)
## 247138) reimbursement2008< 4295 592 344 B1 (0.42 0.41 0.13 0.042 0)
## 494276) age>=82.5 135 71 B1 (0.47 0.36 0.15 0.015 0) *
## 494277) age< 82.5 457 266 B2 (0.4 0.42 0.13 0.05 0)
## 988554) age< 74.5 290 166 B1 (0.43 0.39 0.13 0.052 0)
## 1977108) age>=62.5 234 128 B1 (0.45 0.38 0.12 0.047 0) *
## 1977109) age< 62.5 56 31 B2 (0.32 0.45 0.16 0.071 0) *
## 988555) age>=74.5 167 90 B2 (0.36 0.46 0.13 0.048 0)
## 1977110) reimbursement2008>=4105 39 20 B1 (0.49 0.36 0.13 0.026 0) *
## 1977111) reimbursement2008< 4105 128 65 B2 (0.32 0.49 0.13 0.055 0) *
## 247139) reimbursement2008>=4295 9 1 B2 (0.11 0.89 0 0 0) *
## 61785) reimbursement2008>=4655 719 414 B2 (0.38 0.42 0.15 0.047 0.0042)
## 123570) reimbursement2008< 5835 346 180 B2 (0.35 0.48 0.13 0.038 0.0029) *
## 123571) reimbursement2008>=5835 373 223 B1 (0.4 0.37 0.16 0.056 0.0054)
## 247142) alzheimers>=0.5 124 64 B1 (0.48 0.31 0.15 0.04 0.0081)
## 494284) reimbursement2008< 8555 114 55 B1 (0.52 0.28 0.16 0.035 0.0088) *
## 494285) reimbursement2008>=8555 10 3 B2 (0.1 0.7 0.1 0.1 0) *
## 247143) alzheimers< 0.5 249 149 B2 (0.36 0.4 0.17 0.064 0.004)
## 494286) reimbursement2008>=6045 217 124 B2 (0.36 0.43 0.14 0.069 0.0046) *
## 494287) reimbursement2008< 6045 32 20 B1 (0.38 0.22 0.38 0.031 0)
## 988574) age< 72.5 11 4 B1 (0.64 0.18 0.18 0 0) *
## 988575) age>=72.5 21 11 B3 (0.24 0.24 0.48 0.048 0) *
## 30893) reimbursement2008< 3465 49 24 B2 (0.27 0.51 0.082 0.1 0.041) *
## 15447) age< 50.5 40 26 B1 (0.35 0.3 0.3 0.05 0) *
## 1931) heart.failure>=0.5 2247 1339 B2 (0.35 0.4 0.17 0.066 0.0058)
## 3862) reimbursement2008>=5335 866 530 B1 (0.39 0.37 0.16 0.074 0.0058)
## 7724) reimbursement2008>=8115 129 68 B2 (0.36 0.47 0.12 0.047 0) *
## 7725) reimbursement2008< 8115 737 447 B1 (0.39 0.35 0.17 0.079 0.0068)
## 15450) age< 94.5 703 421 B1 (0.4 0.35 0.17 0.075 0.0071)
## 30900) reimbursement2008>=6635 298 164 B1 (0.45 0.32 0.15 0.067 0.013) *
## 30901) reimbursement2008< 6635 405 255 B2 (0.37 0.37 0.18 0.081 0.0025)
## 61802) reimbursement2008< 5685 137 80 B1 (0.42 0.31 0.16 0.11 0) *
## 61803) reimbursement2008>=5685 268 161 B2 (0.34 0.4 0.19 0.067 0.0037) *
## 15451) age>=94.5 34 17 B2 (0.24 0.5 0.12 0.15 0) *
## 3863) reimbursement2008< 5335 1381 795 B2 (0.33 0.42 0.18 0.061 0.0058)
## 7726) copd< 0.5 997 591 B2 (0.36 0.41 0.17 0.057 0.006)
## 15452) age< 69.5 297 171 B1 (0.42 0.38 0.15 0.04 0.0034)
## 30904) reimbursement2008< 5065 274 153 B1 (0.44 0.36 0.15 0.04 0.0036)
## 61808) alzheimers< 0.5 174 91 B1 (0.48 0.3 0.17 0.046 0.0057) *
## 61809) alzheimers>=0.5 100 54 B2 (0.38 0.46 0.13 0.03 0)
## 123618) reimbursement2008>=4355 26 10 B1 (0.62 0.15 0.15 0.077 0) *
## 123619) reimbursement2008< 4355 74 32 B2 (0.3 0.57 0.12 0.014 0) *
## 30905) reimbursement2008>=5065 23 9 B2 (0.22 0.61 0.13 0.043 0) *
## 15453) age>=69.5 700 407 B2 (0.33 0.42 0.18 0.064 0.0071) *
## 7727) copd>=0.5 384 204 B2 (0.27 0.47 0.19 0.07 0.0052) *
## 483) depression>=0.5 4392 2538 B2 (0.31 0.42 0.18 0.077 0.0061)
## 966) reimbursement2008< 8325 2928 1619 B2 (0.31 0.45 0.17 0.065 0.0065)
## 1932) copd< 0.5 1987 1102 B2 (0.33 0.45 0.16 0.056 0.0045)
## 3864) age< 98.5 1964 1085 B2 (0.33 0.45 0.16 0.057 0.0046)
## 7728) reimbursement2008< 3085 22 8 B1 (0.64 0.23 0.045 0.091 0) *
## 7729) reimbursement2008>=3085 1942 1068 B2 (0.33 0.45 0.16 0.056 0.0046)
## 15458) heart.failure< 0.5 889 508 B2 (0.37 0.43 0.15 0.051 0.0034) *
## 15459) heart.failure>=0.5 1053 560 B2 (0.3 0.47 0.17 0.061 0.0057)
## 30918) osteoporosis< 0.5 721 396 B2 (0.32 0.45 0.16 0.061 0.0055)
## 61836) age>=86.5 109 59 B1 (0.46 0.3 0.15 0.083 0.0092) *
## 61837) age< 86.5 612 320 B2 (0.3 0.48 0.16 0.057 0.0049) *
## 30919) osteoporosis>=0.5 332 164 B2 (0.24 0.51 0.18 0.06 0.006) *
## 3865) age>=98.5 23 12 B3 (0.26 0.26 0.48 0 0) *
## 1933) copd>=0.5 941 517 B2 (0.26 0.45 0.2 0.084 0.011) *
## 967) reimbursement2008>=8325 1464 919 B2 (0.32 0.37 0.2 0.1 0.0055)
## 1934) reimbursement2008< 8485 36 16 B1 (0.56 0.22 0.22 0 0) *
## 1935) reimbursement2008>=8485 1428 891 B2 (0.32 0.38 0.2 0.1 0.0056)
## 3870) age< 78.5 837 532 B2 (0.35 0.36 0.19 0.098 0.0036)
## 7740) reimbursement2008< 21320 639 406 B1 (0.36 0.35 0.19 0.092 0.0031)
## 15480) age< 49.5 83 47 B2 (0.35 0.43 0.096 0.12 0) *
## 15481) age>=49.5 556 352 B1 (0.37 0.33 0.21 0.088 0.0036)
## 30962) age>=67.5 368 230 B1 (0.38 0.36 0.18 0.087 0)
## 61924) reimbursement2008>=10440 261 153 B1 (0.41 0.33 0.17 0.084 0)
## 123848) reimbursement2008< 12585 92 46 B1 (0.5 0.25 0.18 0.065 0) *
## 123849) reimbursement2008>=12585 169 105 B2 (0.37 0.38 0.16 0.095 0)
## 247698) reimbursement2008>=14485 109 63 B1 (0.42 0.31 0.17 0.092 0)
## 495396) osteoporosis< 0.5 72 37 B1 (0.49 0.24 0.17 0.11 0) *
## 495397) osteoporosis>=0.5 37 20 B2 (0.3 0.46 0.19 0.054 0) *
## 247699) reimbursement2008< 14485 60 30 B2 (0.27 0.5 0.13 0.1 0) *
## 61925) reimbursement2008< 10440 107 62 B2 (0.28 0.42 0.21 0.093 0) *
## 30963) age< 67.5 188 122 B1 (0.35 0.28 0.27 0.09 0.011)
## 61926) age>=55.5 135 82 B1 (0.39 0.25 0.27 0.089 0) *
## 61927) age< 55.5 53 34 B2 (0.25 0.36 0.26 0.094 0.038) *
## 7741) reimbursement2008>=21320 198 114 B2 (0.3 0.42 0.16 0.12 0.0051) *
## 3871) age>=78.5 591 359 B2 (0.28 0.39 0.21 0.11 0.0085)
## 7742) heart.failure< 0.5 122 73 B1 (0.4 0.32 0.18 0.098 0)
## 15484) reimbursement2008< 11560 40 17 B1 (0.58 0.2 0.18 0.05 0) *
## 15485) reimbursement2008>=11560 82 51 B2 (0.32 0.38 0.18 0.12 0) *
## 7743) heart.failure>=0.5 469 276 B2 (0.24 0.41 0.22 0.11 0.011) *
## 121) cancer>=0.5 2606 1470 B2 (0.21 0.44 0.25 0.098 0.0081) *
## 61) arthritis>=0.5 9889 5132 B2 (0.22 0.48 0.21 0.088 0.0092)
## 122) depression< 0.5 5134 2665 B2 (0.25 0.48 0.18 0.08 0.0078)
## 244) cancer< 0.5 4305 2260 B2 (0.27 0.48 0.17 0.076 0.0086)
## 488) reimbursement2008>=9880 1063 636 B2 (0.32 0.4 0.18 0.089 0.012)
## 976) ihd< 0.5 102 49 B1 (0.52 0.27 0.13 0.069 0.0098) *
## 977) ihd>=0.5 961 562 B2 (0.29 0.42 0.19 0.092 0.012) *
## 489) reimbursement2008< 9880 3242 1624 B2 (0.25 0.5 0.17 0.072 0.0074) *
## 245) cancer>=0.5 829 405 B2 (0.15 0.51 0.23 0.1 0.0036) *
## 123) depression>=0.5 4755 2467 B2 (0.18 0.48 0.23 0.096 0.011) *
## 31) kidney>=0.5 28734 17139 B2 (0.16 0.4 0.23 0.17 0.03)
## 62) reimbursement2008< 15395 16249 9131 B2 (0.19 0.44 0.24 0.12 0.016)
## 124) arthritis< 0.5 9424 5647 B2 (0.23 0.4 0.23 0.12 0.017)
## 248) cancer< 0.5 7786 4711 B2 (0.25 0.39 0.21 0.12 0.017)
## 496) ihd< 0.5 964 608 B1 (0.37 0.36 0.17 0.085 0.011)
## 992) depression< 0.5 572 338 B1 (0.41 0.32 0.16 0.1 0.01)
## 1984) reimbursement2008< 3545 101 53 B2 (0.33 0.48 0.15 0.04 0.0099) *
## 1985) reimbursement2008>=3545 471 270 B1 (0.43 0.29 0.16 0.11 0.011)
## 3970) osteoporosis< 0.5 346 186 B1 (0.46 0.27 0.15 0.11 0.014) *
## 3971) osteoporosis>=0.5 125 82 B2 (0.33 0.34 0.2 0.13 0)
## 7942) age>=62 106 67 B1 (0.37 0.37 0.15 0.11 0)
## 15884) age>=67.5 93 55 B2 (0.34 0.41 0.16 0.086 0)
## 31768) reimbursement2008>=6110 44 23 B1 (0.48 0.3 0.11 0.11 0)
## 63536) reimbursement2008< 9180 26 9 B1 (0.65 0.15 0.077 0.12 0) *
## 63537) reimbursement2008>=9180 18 9 B2 (0.22 0.5 0.17 0.11 0) *
## 31769) reimbursement2008< 6110 49 24 B2 (0.22 0.51 0.2 0.061 0) *
## 15885) age< 67.5 13 6 B1 (0.54 0.077 0.077 0.31 0) *
## 7943) age< 62 19 10 B3 (0.11 0.21 0.47 0.21 0) *
## 993) depression>=0.5 392 227 B2 (0.31 0.42 0.19 0.064 0.013)
## 1986) reimbursement2008>=14460 9 2 B1 (0.78 0.22 0 0 0) *
## 1987) reimbursement2008< 14460 383 220 B2 (0.3 0.43 0.2 0.065 0.013) *
## 497) ihd>=0.5 6822 4095 B2 (0.24 0.4 0.22 0.12 0.018)
## 994) reimbursement2008< 6325 3172 1786 B2 (0.22 0.44 0.22 0.11 0.016) *
## 995) reimbursement2008>=6325 3650 2309 B2 (0.25 0.37 0.22 0.14 0.019)
## 1990) osteoporosis< 0.5 2424 1594 B2 (0.27 0.34 0.23 0.14 0.02)
## 3980) depression< 0.5 1234 816 B2 (0.3 0.34 0.2 0.13 0.024)
## 7960) reimbursement2008>=12135 349 226 B1 (0.35 0.3 0.16 0.16 0.032)
## 15920) age>=54 331 210 B1 (0.37 0.28 0.16 0.16 0.03) *
## 15921) age< 54 18 9 B2 (0.11 0.5 0.11 0.22 0.056) *
## 7961) reimbursement2008< 12135 885 570 B2 (0.28 0.36 0.22 0.12 0.021) *
## 3981) depression>=0.5 1190 778 B2 (0.24 0.35 0.25 0.15 0.016)
## 7962) copd< 0.5 547 367 B2 (0.28 0.33 0.25 0.12 0.022)
## 15924) reimbursement2008>=9205 310 209 B1 (0.33 0.32 0.21 0.12 0.029)
## 31848) reimbursement2008< 9955 50 28 B2 (0.42 0.44 0.02 0.12 0) *
## 31849) reimbursement2008>=9955 260 180 B1 (0.31 0.3 0.24 0.12 0.035)
## 63698) reimbursement2008>=14765 20 9 B1 (0.55 0.25 0.1 0.05 0.05) *
## 63699) reimbursement2008< 14765 240 168 B2 (0.29 0.3 0.25 0.12 0.033)
## 127398) age>=61.5 201 138 B1 (0.31 0.29 0.24 0.13 0.02)
## 254796) reimbursement2008< 12625 112 69 B1 (0.38 0.24 0.28 0.089 0.0089) *
## 254797) reimbursement2008>=12625 89 58 B2 (0.22 0.35 0.2 0.19 0.034) *
## 127399) age< 61.5 39 25 B2 (0.15 0.36 0.31 0.077 0.1) *
## 15925) reimbursement2008< 9205 237 156 B2 (0.22 0.34 0.3 0.12 0.013)
## 31850) age< 67.5 56 29 B2 (0.2 0.48 0.16 0.16 0) *
## 31851) age>=67.5 181 118 B3 (0.23 0.3 0.35 0.1 0.017)
## 63702) reimbursement2008>=6865 136 82 B3 (0.25 0.26 0.4 0.074 0.015) *
## 63703) reimbursement2008< 6865 45 27 B2 (0.18 0.4 0.2 0.2 0.022) *
## 7963) copd>=0.5 643 411 B2 (0.21 0.36 0.25 0.17 0.011) *
## 1991) osteoporosis>=0.5 1226 715 B2 (0.21 0.42 0.22 0.14 0.017) *
## 249) cancer>=0.5 1638 936 B2 (0.13 0.43 0.29 0.14 0.016) *
## 125) arthritis>=0.5 6825 3484 B2 (0.13 0.49 0.25 0.12 0.014) *
## 63) reimbursement2008>=15395 12485 8008 B2 (0.13 0.36 0.23 0.24 0.049)
## 126) arthritis>=0.5 5402 3220 B2 (0.094 0.4 0.24 0.22 0.04)
## 252) reimbursement2008< 34925 3345 1942 B2 (0.11 0.42 0.25 0.19 0.03)
## 504) depression< 0.5 1291 714 B2 (0.14 0.45 0.22 0.17 0.025)
## 1008) cancer< 0.5 973 546 B2 (0.16 0.44 0.19 0.18 0.029) *
## 1009) cancer>=0.5 318 168 B2 (0.072 0.47 0.3 0.14 0.013)
## 2018) reimbursement2008>=16525 293 144 B2 (0.068 0.51 0.28 0.13 0.014) *
## 2019) reimbursement2008< 16525 25 10 B3 (0.12 0.04 0.6 0.24 0) *
## 505) depression>=0.5 2054 1228 B2 (0.092 0.4 0.27 0.2 0.034) *
## 253) reimbursement2008>=34925 2057 1278 B2 (0.067 0.38 0.23 0.27 0.055)
## 506) copd< 0.5 520 300 B2 (0.096 0.42 0.23 0.21 0.042) *
## 507) copd>=0.5 1537 978 B2 (0.057 0.36 0.23 0.29 0.06)
## 1014) age>=62.5 1286 804 B2 (0.058 0.37 0.24 0.27 0.061) *
## 1015) age< 62.5 251 153 B4 (0.052 0.31 0.2 0.39 0.052)
## 2030) reimbursement2008< 101585 237 150 B4 (0.055 0.32 0.21 0.37 0.051)
## 4060) cancer>=0.5 62 36 B2 (0.048 0.42 0.27 0.24 0.016) *
## 4061) cancer< 0.5 175 103 B4 (0.057 0.29 0.18 0.41 0.063) *
## 2031) reimbursement2008>=101585 14 3 B4 (0 0.071 0.071 0.79 0.071) *
## 127) arthritis< 0.5 7083 4788 B2 (0.15 0.32 0.22 0.25 0.057)
## 254) cancer< 0.5 5298 3651 B2 (0.17 0.31 0.2 0.26 0.062)
## 508) depression< 0.5 2489 1797 B2 (0.22 0.28 0.18 0.27 0.06)
## 1016) copd>=0.5 1317 890 B2 (0.2 0.32 0.18 0.24 0.056)
## 2032) ihd< 0.5 72 41 B1 (0.43 0.25 0.15 0.11 0.056) *
## 2033) ihd>=0.5 1245 836 B2 (0.19 0.33 0.18 0.24 0.056) *
## 1017) copd< 0.5 1172 815 B4 (0.23 0.23 0.17 0.3 0.065)
## 2034) reimbursement2008>=43640 191 129 B2 (0.15 0.32 0.23 0.22 0.073)
## 4068) age>=64.5 172 112 B2 (0.16 0.35 0.2 0.23 0.064) *
## 4069) age< 64.5 19 10 B3 (0.11 0.11 0.47 0.16 0.16) *
## 2035) reimbursement2008< 43640 981 666 B4 (0.25 0.21 0.16 0.32 0.063)
## 4070) reimbursement2008< 23175 468 337 B1 (0.28 0.24 0.16 0.26 0.056)
## 8140) age< 93.5 457 326 B1 (0.29 0.23 0.16 0.26 0.057)
## 16280) age< 86.5 398 288 B1 (0.28 0.25 0.16 0.25 0.063)
## 32560) alzheimers>=0.5 179 122 B1 (0.32 0.28 0.15 0.17 0.073)
## 65120) reimbursement2008>=21440 38 19 B2 (0.24 0.5 0.11 0.13 0.026) *
## 65121) reimbursement2008< 21440 141 93 B1 (0.34 0.23 0.16 0.18 0.085)
## 130242) reimbursement2008>=17585 89 53 B1 (0.4 0.17 0.19 0.16 0.079) *
## 130243) reimbursement2008< 17585 52 35 B2 (0.23 0.33 0.12 0.23 0.096) *
## 32561) alzheimers< 0.5 219 151 B4 (0.24 0.23 0.16 0.31 0.055) *
## 16281) age>=86.5 59 38 B1 (0.36 0.1 0.17 0.36 0.017)
## 32562) reimbursement2008>=19680 18 8 B1 (0.56 0.11 0 0.28 0.056) *
## 32563) reimbursement2008< 19680 41 25 B4 (0.27 0.098 0.24 0.39 0) *
## 8141) age>=93.5 11 6 B2 (0 0.45 0.36 0.18 0) *
## 4071) reimbursement2008>=23175 513 320 B4 (0.22 0.18 0.16 0.38 0.07) *
## 509) depression>=0.5 2809 1854 B2 (0.13 0.34 0.22 0.25 0.063) *
## 255) cancer>=0.5 1785 1137 B2 (0.097 0.36 0.28 0.22 0.041) *## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 171981 12159 140 186 0
## B2 26534 25373 156 196 0
## B3 12021 12119 286 160 0
## B4 4652 6824 70 360 0
## B5 483 1029 14 60 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7205162 NA 0.7188342 0.7221934 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 0.0000000## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114141 8610 124 103 0
## B2 18409 16102 187 142 0
## B3 8027 8146 118 99 0
## B4 3099 4584 53 201 0
## B5 351 657 4 45 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.126669e-01 NA 7.105887e-01 7.147384e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 7.015238e-319 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## model_id model_method
## 1 All.X.lser.ys.cp.4015.rpart rpart
## feats
## 1 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 1 75.838 16.151
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.7205162 0.7188342 0.7221934
## max.Kappa.fit min.loss.error.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.7582887 0.7126669 0.7105887
## max.AccuracyUpper.OOB max.Kappa.OOB min.loss.error.OOB min.SSE.fit
## 1 0.7147384 NA 0.7578902 0
## min.loss.errorSD.fit
## 1 0.005993538
## [1] "iterating over method:rf"
## [1] "fitting model: All.X.lser.no.cp.opt.rf"
## [1] " indep_vars: age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm"## Loading required package: randomForest
## randomForest 4.6-10
## Type rfNews() to see new features/changes/bug fixes.
## + : mtry= 2
## - : mtry= 2
## + : mtry= 8
## - : mtry= 8
## + : mtry=15
## - : mtry=15
## Aggregating results
## Selecting tuning parameters
## Fitting mtry = 2 on full training set## Warning in myfit_mdl_fn(model_id = paste0(model_id_pfx,
## ".lser.no.cp.opt"), : model's bestTune found at an extreme of tuneGrid for
## parameter: mtry
## Length Class Mode
## call 4 -none- call
## type 1 -none- character
## predicted 274803 factor numeric
## err.rate 3000 -none- numeric
## confusion 30 -none- numeric
## votes 1374015 matrix numeric
## oob.times 274803 -none- numeric
## classes 5 -none- character
## importance 15 -none- numeric
## importanceSD 0 -none- NULL
## localImportance 0 -none- NULL
## proximity 0 -none- NULL
## ntree 1 -none- numeric
## mtry 1 -none- numeric
## forest 14 -none- list
## y 274803 factor numeric
## test 0 -none- NULL
## inbag 0 -none- NULL
## xNames 15 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 5 -none- character## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 173343 11123 0 0 0
## B2 27770 24489 0 0 0
## B3 12454 12124 8 0 0
## B4 4772 7134 0 0 0
## B5 497 1089 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7199339 NA 0.7182509 0.7216123 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 115110 7868 0 0 0
## B2 19054 15786 0 0 0
## B3 8252 8138 0 0 0
## B4 3213 4724 0 0 0
## B5 357 700 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7144900 NA 0.7124157 0.7165575 0.6712700
## AccuracyPValue McnemarPValue
## 0.0000000 NaN
## model_id model_method
## 1 All.X.lser.no.cp.opt.rf rf
## feats
## 1 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 3 3479.97 449.343
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.7139405 0.7182509 0.7216123
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 0.3281909 0.71449 0.7124157
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit
## 1 0.7165575 NA 0
## [1] "fitting model: All.X.lser.ys.cp.opt.rf"
## [1] " indep_vars: age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm"## Warning in train.default(x, y, weights = w, ...): The metric "loss.error"
## was not in the result set. Accuracy will be used instead.
## + : mtry= 2
## - : mtry= 2
## + : mtry= 8
## - : mtry= 8
## + : mtry=15
## - : mtry=15
## Aggregating results
## Selecting tuning parameters
## Fitting mtry = 15 on full training set## Warning in myfit_mdl_fn(model_id = paste0(model_id_pfx,
## ".lser.ys.cp.opt"), : model's bestTune found at an extreme of tuneGrid for
## parameter: mtry
## Length Class Mode
## call 4 -none- call
## type 1 -none- character
## predicted 274803 factor numeric
## err.rate 3000 -none- numeric
## confusion 30 -none- numeric
## votes 1374015 matrix numeric
## oob.times 274803 -none- numeric
## classes 5 -none- character
## importance 15 -none- numeric
## importanceSD 0 -none- NULL
## localImportance 0 -none- NULL
## proximity 0 -none- NULL
## ntree 1 -none- numeric
## mtry 1 -none- numeric
## forest 14 -none- list
## y 274803 factor numeric
## test 0 -none- NULL
## inbag 0 -none- NULL
## xNames 15 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 5 -none- character## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 184466 0 0 0 0
## B2 0 52259 0 0 0
## B3 0 0 24586 0 0
## B4 0 0 0 11906 0
## B5 0 0 0 0 1586
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 1.0000000 NA 0.9999866 1.0000000 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 111655 9615 1279 416 13
## B2 18013 13796 2216 796 19
## B3 8029 6719 1167 465 10
## B4 3148 3538 718 513 20
## B5 360 458 122 114 3
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.939553e-01 NA 6.918392e-01 6.960652e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 2.423558e-96 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow
## model_id model_method
## 1 All.X.lser.ys.cp.opt.rf rf
## feats
## 1 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 3 4381.707 1409.455
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 1 0.9999866 1
## max.Kappa.fit min.loss.error.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0 0.6939553 0.6918392
## max.AccuracyUpper.OOB max.Kappa.OOB min.loss.error.OOB min.SSE.fit
## 1 0.6960652 NA 0.7655102 0
## min.Accuracy.fit min.Kappa.fit
## 1 0.6943265 0.3128527# Simplify a model
# fit_df <- glb_entity_df; glb_mdl <- step(<complex>_mdl)
print(glb_models_df)## model_id model_method
## 1 Baseline.mybaseln_classfr mybaseln_classfr
## 2 MFO.myMFO_classfr myMFO_classfr
## 3 Random.myrandom_classfr myrandom_classfr
## 4 Max.cor.Y.cv.0.rpart rpart
## 5 Max.cor.Y.cv.G.rpart rpart
## 6 Interact.High.cor.y.rpart rpart
## 7 Low.cor.X.rpart rpart
## 8 All.X.lser.no.cp.opt.rpart rpart
## 9 All.X.lser.no.cp.4015.rpart rpart
## 10 All.X.lser.ys.cp.opt.rpart rpart
## 11 All.X.lser.ys.cp.4015.rpart rpart
## 12 All.X.lser.no.cp.opt.rf rf
## 13 All.X.lser.ys.cp.opt.rf rf
## feats
## 1 bucket2008.fctr, .rnorm
## 2 .rnorm
## 3 .rnorm
## 4 bucket2008
## 5 bucket2008
## 6 bucket2008, reimbursement2008
## 7 bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age
## 8 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 9 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 10 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 11 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 12 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## 13 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 0 0.626 0.006
## 2 0 0.515 0.038
## 3 0 0.375 0.035
## 4 0 4.811 3.430
## 5 3 20.801 3.333
## 6 3 31.301 5.050
## 7 3 101.182 13.937
## 8 3 107.610 15.544
## 9 1 73.967 15.630
## 10 3 108.644 15.505
## 11 1 75.838 16.151
## 12 3 3479.970 449.343
## 13 3 4381.707 1409.455
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.6829729 0.6812294 0.6847125
## 2 0.6712663 0.6695064 0.6730227
## 3 0.4967267 0.4948556 0.4985980
## 4 0.6712663 0.6695064 0.6730227
## 5 0.7031401 0.7014279 0.7048479
## 6 0.7029800 0.7018289 0.7052475
## 7 0.7086058 0.7063706 0.7097740
## 8 0.7086604 0.7064253 0.7098285
## 9 0.7117062 0.7188342 0.7221934
## 10 0.7035404 0.7018289 0.7052475
## 11 0.7205162 0.7188342 0.7221934
## 12 0.7139405 0.7182509 0.7216123
## 13 1.0000000 0.9999866 1.0000000
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.6838135 0.6816786
## 2 NA 0.6712700 0.6691135
## 3 NA 0.4948800 0.4925879
## 4 NA 0.6712700 0.6691135
## 5 0.3440289 0.7040207 0.7019245
## 6 0.3368565 0.7039770 0.7018807
## 7 0.3161651 0.7077434 0.7056548
## 8 0.3140169 0.7074814 0.7053922
## 9 0.3328722 0.7126669 0.7105887
## 10 NA 0.7039770 0.7018807
## 11 NA 0.7126669 0.7105887
## 12 0.3281909 0.7144900 0.7124157
## 13 NA 0.6939553 0.6918392
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.6859425 NA 0 NA
## 2 0.6734210 NA 0 NA
## 3 0.4971722 NA 0 NA
## 4 0.6734210 NA 0 NA
## 5 0.7061105 NA 0 0.0016657885
## 6 0.7060669 NA 0 0.0016900317
## 7 0.7098254 NA 0 0.0009742811
## 8 0.7095639 NA 0 0.0010140545
## 9 0.7147384 NA 0 0.0021496212
## 10 0.7060669 NA 0 NA
## 11 0.7147384 NA 0 NA
## 12 0.7165575 NA 0 NA
## 13 0.6960652 NA 0 NA
## max.KappaSD.fit min.loss.error.fit min.loss.error.OOB
## 1 NA NA NA
## 2 NA NA NA
## 3 NA NA NA
## 4 NA NA NA
## 5 0.004508795 NA NA
## 6 0.006361514 NA NA
## 7 0.004868262 NA NA
## 8 0.004809558 NA NA
## 9 0.007338010 NA NA
## 10 NA 0.7784778 0.7441403
## 11 NA 0.7582887 0.7578902
## 12 NA NA NA
## 13 NA 0.0000000 0.7655102
## min.loss.errorSD.fit min.Accuracy.fit min.Kappa.fit
## 1 NA NA NA
## 2 NA NA NA
## 3 NA NA NA
## 4 NA NA NA
## 5 NA NA NA
## 6 NA NA NA
## 7 NA NA NA
## 8 NA NA NA
## 9 NA NA NA
## 10 0.016425281 NA NA
## 11 0.005993538 NA NA
## 12 NA NA NA
## 13 NA 0.6943265 0.3128527if (!is.null(glb_model_metric_smmry)) {
stats_df <- glb_models_df[, "model_id", FALSE]
stats_mdl_df <- data.frame()
for (model_id in stats_df$model_id) {
stats_mdl_df <- rbind(stats_mdl_df,
mypredict_mdl(glb_models_lst[[model_id]], glb_entity_df, glb_rsp_var,
glb_rsp_var_out, model_id, "fit",
glb_model_metric_smmry, glb_model_metric,
glb_model_metric_maximize, ret_type="stats"))
}
stats_df <- merge(stats_df, stats_mdl_df, all.x=TRUE)
stats_mdl_df <- data.frame()
for (model_id in stats_df$model_id) {
stats_mdl_df <- rbind(stats_mdl_df,
mypredict_mdl(glb_models_lst[[model_id]], glb_newent_df, glb_rsp_var,
glb_rsp_var_out, model_id, "OOB",
glb_model_metric_smmry, glb_model_metric,
glb_model_metric_maximize, ret_type="stats"))
}
stats_df <- merge(stats_df, stats_mdl_df, all.x=TRUE)
# tmp_models_df <- orderBy(~model_id, glb_models_df)
# rownames(tmp_models_df) <- seq(1, nrow(tmp_models_df))
# all.equal(subset(tmp_models_df[, names(stats_df)], model_id != "Random.myrandom_classfr"),
# subset(stats_df, model_id != "Random.myrandom_classfr"))
# print(subset(tmp_models_df[, names(stats_df)], model_id != "Random.myrandom_classfr")[, c("model_id", "max.Accuracy.fit")])
# print(subset(stats_df, model_id != "Random.myrandom_classfr")[, c("model_id", "max.Accuracy.fit")])
print("Merging following data into glb_models_df:")
print(stats_mrg_df <- stats_df[, c(1, grep(glb_model_metric, names(stats_df)))])
print(tmp_models_df <- orderBy(~model_id, glb_models_df[, c("model_id", grep(glb_model_metric, names(stats_df), value=TRUE))]))
tmp2_models_df <- glb_models_df[, c("model_id", setdiff(names(glb_models_df), grep(glb_model_metric, names(stats_df), value=TRUE)))]
tmp3_models_df <- merge(tmp2_models_df, stats_mrg_df, all.x=TRUE, sort=FALSE)
print(tmp3_models_df)
print(names(tmp3_models_df))
print(glb_models_df <- subset(tmp3_models_df, select=-model_id.1))
}## [1] "in Baseline.Classifier$predict"
## [1] "class(newdata):"
## [1] "matrix"
## [1] "head(newdata):"
## bucket2008.fctrB2 bucket2008.fctrB3 bucket2008.fctrB4 bucket2008.fctrB5
## 1 0 0 0 0
## 2 0 0 0 0
## 4 0 0 0 0
## 7 0 0 0 0
## 11 0 0 0 0
## 13 0 0 0 0
## .rnorm
## 1 -0.9248001
## 2 0.1533902
## 4 -0.9917587
## 7 -0.5160354
## 11 -1.0376029
## 13 -0.3801495
## [1] "x_names: "
## [1] "bucket2008.fctrB2" "bucket2008.fctrB3" "bucket2008.fctrB4"
## [4] "bucket2008.fctrB5"
## [1] "x_vals: "
## [1] "B1" "B2" "B3" "B4" "B5"
## [1] "length(y):"
## [1] 274803
## [1] "head(y):"
## [1] B1 B1 B1 B1 B1 B1
## Levels: B1 B2 B3 B4 B5## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 164967 11747 5214 2275 263
## B2 24001 16172 6877 4367 842
## B3 10679 6848 4004 2552 503
## B4 4020 2835 2081 2399 571
## B5 410 300 266 469 141
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.829729e-01 NA 6.812294e-01 6.847125e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 1.601032e-39 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## [1] "in MFO.Classifier$predict"## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 184466 0 0 0 0
## B2 52259 0 0 0 0
## B3 24586 0 0 0 0
## B4 11906 0 0 0 0
## B5 1586 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712663 NA 0.6695064 0.6730227 0.6712663
## AccuracyPValue McnemarPValue
## 0.5009025 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 123803 35402 16329 7862 1070
## B2 35169 9822 4708 2274 286
## B3 16479 4731 2205 1036 135
## B4 7964 2307 1095 481 59
## B5 1041 324 143 71 7
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.4960572 NA 0.4941860 0.4979284 0.6712663
## AccuracyPValue McnemarPValue
## 1.0000000 0.6148447## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 184466 0 0 0 0
## B2 52259 0 0 0 0
## B3 24586 0 0 0 0
## B4 11906 0 0 0 0
## B5 1586 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712663 NA 0.6695064 0.6730227 0.6712663
## AccuracyPValue McnemarPValue
## 0.5009025 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 164967 19499 0 0 0
## B2 24001 28258 0 0 0
## B3 10679 13907 0 0 0
## B4 4020 7886 0 0 0
## B5 410 1176 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.031401e-01 NA 7.014279e-01 7.048479e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 3.041172e-282 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 165494 18972 0 0 0
## B2 24418 27841 0 0 0
## B3 10858 13728 0 0 0
## B4 4097 7809 0 0 0
## B5 418 1168 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.035404e-01 NA 7.018289e-01 7.052475e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 2.207353e-289 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 171934 12532 0 0 0
## B2 29612 22647 0 0 0
## B3 13099 11487 0 0 0
## B4 5037 6869 0 0 0
## B5 523 1063 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7080745 NA 0.7063706 0.7097740 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 172245 12221 0 0 0
## B2 29908 22351 0 0 0
## B3 13223 11363 0 0 0
## B4 5096 6810 0 0 0
## B5 530 1056 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7081291 NA 0.7064253 0.7098285 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 171981 12159 140 186 0
## B2 26534 25373 156 196 0
## B3 12021 12119 286 160 0
## B4 4652 6824 70 360 0
## B5 483 1029 14 60 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7205162 NA 0.7188342 0.7221934 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 0.0000000## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 165494 18972 0 0 0
## B2 24418 27841 0 0 0
## B3 10858 13728 0 0 0
## B4 4097 7809 0 0 0
## B5 418 1168 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.035404e-01 NA 7.018289e-01 7.052475e-01 6.712663e-01
## AccuracyPValue McnemarPValue
## 2.207353e-289 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 171981 12159 140 186 0
## B2 26534 25373 156 196 0
## B3 12021 12119 286 160 0
## B4 4652 6824 70 360 0
## B5 483 1029 14 60 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7205162 NA 0.7188342 0.7221934 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 0.0000000## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 173342 11124 0 0 0
## B2 27769 24490 0 0 0
## B3 12454 12124 8 0 0
## B4 4772 7134 0 0 0
## B5 497 1089 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7199339 NA 0.7182509 0.7216123 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 184466 0 0 0 0
## B2 0 52259 0 0 0
## B3 0 0 24586 0 0
## B4 0 0 0 11906 0
## B5 0 0 0 0 1586
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 1.0000000 NA 0.9999866 1.0000000 0.6712663
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114141 8610 124 103 0
## B2 18409 16102 187 142 0
## B3 8027 8146 118 99 0
## B4 3099 4584 53 201 0
## B5 351 657 4 45 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.126669e-01 NA 7.105887e-01 7.147384e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 7.015238e-319 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 115108 7870 0 0 0
## B2 19053 15787 0 0 0
## B3 8254 8136 0 0 0
## B4 3211 4726 0 0 0
## B5 357 700 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.7144846 NA 0.7124102 0.7165521 0.6712700
## AccuracyPValue McnemarPValue
## 0.0000000 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114902 8076 0 0 0
## B2 20130 14710 0 0 0
## B3 8749 7641 0 0 0
## B4 3409 4528 0 0 0
## B5 382 675 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.074814e-01 NA 7.053922e-01 7.095639e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 8.205962e-244 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114141 8610 124 103 0
## B2 18409 16102 187 142 0
## B3 8027 8146 118 99 0
## B4 3099 4584 53 201 0
## B5 351 657 4 45 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.126669e-01 NA 7.105887e-01 7.147384e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 7.015238e-319 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 111666 9606 1277 416 13
## B2 18015 13793 2212 801 19
## B3 8026 6727 1164 463 10
## B4 3149 3541 715 512 20
## B5 361 457 121 115 3
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.939771e-01 NA 6.918611e-01 6.960870e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 1.592264e-96 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110452 12526 0 0 0
## B2 16322 18518 0 0 0
## B3 7105 9285 0 0 0
## B4 2740 5197 0 0 0
## B5 299 758 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.039770e-01 NA 7.018807e-01 7.060669e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 6.148674e-199 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## [1] "in Baseline.Classifier$predict"
## [1] "class(newdata):"
## [1] "matrix"
## [1] "head(newdata):"
## bucket2008.fctrB2 bucket2008.fctrB3 bucket2008.fctrB4 bucket2008.fctrB5
## 3 0 0 0 0
## 5 0 0 0 0
## 6 0 0 0 0
## 8 0 0 0 0
## 9 0 0 0 0
## 10 0 0 0 0
## .rnorm
## 3 0.7183313438
## 5 0.0008759006
## 6 -1.4428971189
## 8 -1.9680342619
## 9 -3.0026827087
## 10 -1.1723168041
## [1] "x_names: "
## [1] "bucket2008.fctrB2" "bucket2008.fctrB3" "bucket2008.fctrB4"
## [4] "bucket2008.fctrB5"
## [1] "x_vals: "
## [1] "B1" "B2" "B3" "B4" "B5"
## [1] "length(y):"
## [1] 183202
## [1] "head(y):"
## [1] B1 B1 B1 B1 B1 B1
## Levels: B1 B2 B3 B4 B5## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110138 7787 3427 1452 174
## B2 16000 10721 4629 2931 559
## B3 7006 4629 2774 1621 360
## B4 2688 1943 1415 1539 352
## B5 293 191 160 309 104
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 6.838135e-01 NA 6.816786e-01 6.859425e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 9.943570e-31 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110452 12526 0 0 0
## B2 16322 18518 0 0 0
## B3 7105 9285 0 0 0
## B4 2740 5197 0 0 0
## B5 299 758 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.039770e-01 NA 7.018807e-01 7.060669e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 6.148674e-199 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114716 8262 0 0 0
## B2 19896 14944 0 0 0
## B3 8672 7718 0 0 0
## B4 3366 4571 0 0 0
## B5 379 678 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.077434e-01 NA 7.056548e-01 7.098254e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 2.343725e-247 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 122978 0 0 0 0
## B2 34840 0 0 0 0
## B3 16390 0 0 0 0
## B4 7937 0 0 0 0
## B5 1057 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712700 NA 0.6691135 0.6734210 0.6712700
## AccuracyPValue McnemarPValue
## 0.5011054 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 110138 12840 0 0 0
## B2 16000 18840 0 0 0
## B3 7006 9384 0 0 0
## B4 2688 5249 0 0 0
## B5 293 764 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.040207e-01 NA 7.019245e-01 7.061105e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 1.813355e-199 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## [1] "in MFO.Classifier$predict"## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 122978 0 0 0 0
## B2 34840 0 0 0 0
## B3 16390 0 0 0 0
## B4 7937 0 0 0 0
## B5 1057 0 0 0 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.6712700 NA 0.6691135 0.6734210 0.6712700
## AccuracyPValue McnemarPValue
## 0.5011054 NaN## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 82872 23225 10833 5356 692
## B2 23229 6624 3189 1594 204
## B3 10992 3080 1504 735 79
## B4 5323 1542 704 328 40
## B5 674 219 112 45 7
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 0.4985481 NA 0.4962558 0.5008403 0.6712700
## AccuracyPValue McnemarPValue
## 1.0000000 0.3234960## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## [1] "Merging following data into glb_models_df:"
## model_id min.loss.error.fit min.loss.error.OOB
## 1 All.X.lser.no.cp.4015.rpart 0.7447408 0.7578902
## 2 All.X.lser.no.cp.opt.rf 0.7583760 0.7668584
## 3 All.X.lser.no.cp.opt.rpart 0.7861887 0.7875787
## 4 All.X.lser.ys.cp.4015.rpart 0.7447408 0.7578902
## 5 All.X.lser.ys.cp.opt.rf 0.0000000 0.7655702
## 6 All.X.lser.ys.cp.opt.rpart 0.7455013 0.7441403
## 7 Baseline.mybaseln_classfr 0.7402976 0.7386055
## 8 Interact.High.cor.y.rpart 0.7455013 0.7441403
## 9 Low.cor.X.rpart 0.7837833 0.7846967
## 10 Max.cor.Y.cv.0.rpart 1.0443336 1.0443008
## 11 Max.cor.Y.cv.G.rpart 0.7424628 0.7406251
## 12 MFO.myMFO_classfr 1.0443336 1.0443008
## 13 Random.myrandom_classfr 1.1765119 1.1735298
## model_id min.loss.error.fit min.loss.error.OOB
## 9 All.X.lser.no.cp.4015.rpart NA NA
## 12 All.X.lser.no.cp.opt.rf NA NA
## 8 All.X.lser.no.cp.opt.rpart NA NA
## 11 All.X.lser.ys.cp.4015.rpart 0.7582887 0.7578902
## 13 All.X.lser.ys.cp.opt.rf 0.0000000 0.7655102
## 10 All.X.lser.ys.cp.opt.rpart 0.7784778 0.7441403
## 1 Baseline.mybaseln_classfr NA NA
## 6 Interact.High.cor.y.rpart NA NA
## 7 Low.cor.X.rpart NA NA
## 4 Max.cor.Y.cv.0.rpart NA NA
## 5 Max.cor.Y.cv.G.rpart NA NA
## 2 MFO.myMFO_classfr NA NA
## 3 Random.myrandom_classfr NA NA
## model_id model_id.1
## 1 Baseline.mybaseln_classfr Baseline.mybaseln_classfr
## 2 MFO.myMFO_classfr MFO.myMFO_classfr
## 3 Random.myrandom_classfr Random.myrandom_classfr
## 4 Max.cor.Y.cv.0.rpart Max.cor.Y.cv.0.rpart
## 5 Max.cor.Y.cv.G.rpart Max.cor.Y.cv.G.rpart
## 6 Interact.High.cor.y.rpart Interact.High.cor.y.rpart
## 7 Low.cor.X.rpart Low.cor.X.rpart
## 8 All.X.lser.no.cp.opt.rpart All.X.lser.no.cp.opt.rpart
## 9 All.X.lser.no.cp.4015.rpart All.X.lser.no.cp.4015.rpart
## 10 All.X.lser.ys.cp.opt.rpart All.X.lser.ys.cp.opt.rpart
## 11 All.X.lser.ys.cp.4015.rpart All.X.lser.ys.cp.4015.rpart
## 12 All.X.lser.no.cp.opt.rf All.X.lser.no.cp.opt.rf
## 13 All.X.lser.ys.cp.opt.rf All.X.lser.ys.cp.opt.rf
## model_method
## 1 mybaseln_classfr
## 2 myMFO_classfr
## 3 myrandom_classfr
## 4 rpart
## 5 rpart
## 6 rpart
## 7 rpart
## 8 rpart
## 9 rpart
## 10 rpart
## 11 rpart
## 12 rf
## 13 rf
## feats
## 1 bucket2008.fctr, .rnorm
## 2 .rnorm
## 3 .rnorm
## 4 bucket2008
## 5 bucket2008
## 6 bucket2008, reimbursement2008
## 7 bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age
## 8 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 9 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 10 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 11 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 12 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## 13 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 0 0.626 0.006
## 2 0 0.515 0.038
## 3 0 0.375 0.035
## 4 0 4.811 3.430
## 5 3 20.801 3.333
## 6 3 31.301 5.050
## 7 3 101.182 13.937
## 8 3 107.610 15.544
## 9 1 73.967 15.630
## 10 3 108.644 15.505
## 11 1 75.838 16.151
## 12 3 3479.970 449.343
## 13 3 4381.707 1409.455
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.6829729 0.6812294 0.6847125
## 2 0.6712663 0.6695064 0.6730227
## 3 0.4967267 0.4948556 0.4985980
## 4 0.6712663 0.6695064 0.6730227
## 5 0.7031401 0.7014279 0.7048479
## 6 0.7029800 0.7018289 0.7052475
## 7 0.7086058 0.7063706 0.7097740
## 8 0.7086604 0.7064253 0.7098285
## 9 0.7117062 0.7188342 0.7221934
## 10 0.7035404 0.7018289 0.7052475
## 11 0.7205162 0.7188342 0.7221934
## 12 0.7139405 0.7182509 0.7216123
## 13 1.0000000 0.9999866 1.0000000
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.6838135 0.6816786
## 2 NA 0.6712700 0.6691135
## 3 NA 0.4948800 0.4925879
## 4 NA 0.6712700 0.6691135
## 5 0.3440289 0.7040207 0.7019245
## 6 0.3368565 0.7039770 0.7018807
## 7 0.3161651 0.7077434 0.7056548
## 8 0.3140169 0.7074814 0.7053922
## 9 0.3328722 0.7126669 0.7105887
## 10 NA 0.7039770 0.7018807
## 11 NA 0.7126669 0.7105887
## 12 0.3281909 0.7144900 0.7124157
## 13 NA 0.6939553 0.6918392
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.6859425 NA 0 NA
## 2 0.6734210 NA 0 NA
## 3 0.4971722 NA 0 NA
## 4 0.6734210 NA 0 NA
## 5 0.7061105 NA 0 0.0016657885
## 6 0.7060669 NA 0 0.0016900317
## 7 0.7098254 NA 0 0.0009742811
## 8 0.7095639 NA 0 0.0010140545
## 9 0.7147384 NA 0 0.0021496212
## 10 0.7060669 NA 0 NA
## 11 0.7147384 NA 0 NA
## 12 0.7165575 NA 0 NA
## 13 0.6960652 NA 0 NA
## max.KappaSD.fit min.loss.errorSD.fit min.Accuracy.fit min.Kappa.fit
## 1 NA NA NA NA
## 2 NA NA NA NA
## 3 NA NA NA NA
## 4 NA NA NA NA
## 5 0.004508795 NA NA NA
## 6 0.006361514 NA NA NA
## 7 0.004868262 NA NA NA
## 8 0.004809558 NA NA NA
## 9 0.007338010 NA NA NA
## 10 NA 0.016425281 NA NA
## 11 NA 0.005993538 NA NA
## 12 NA NA NA NA
## 13 NA NA 0.6943265 0.3128527
## min.loss.error.fit min.loss.error.OOB
## 1 0.7402976 0.7386055
## 2 1.0443336 1.0443008
## 3 1.1765119 1.1735298
## 4 1.0443336 1.0443008
## 5 0.7424628 0.7406251
## 6 0.7455013 0.7441403
## 7 0.7837833 0.7846967
## 8 0.7861887 0.7875787
## 9 0.7447408 0.7578902
## 10 0.7455013 0.7441403
## 11 0.7447408 0.7578902
## 12 0.7583760 0.7668584
## 13 0.0000000 0.7655702
## [1] "model_id" "model_id.1"
## [3] "model_method" "feats"
## [5] "max.nTuningRuns" "min.elapsedtime.everything"
## [7] "min.elapsedtime.final" "max.Accuracy.fit"
## [9] "max.AccuracyLower.fit" "max.AccuracyUpper.fit"
## [11] "max.Kappa.fit" "max.Accuracy.OOB"
## [13] "max.AccuracyLower.OOB" "max.AccuracyUpper.OOB"
## [15] "max.Kappa.OOB" "min.SSE.fit"
## [17] "max.AccuracySD.fit" "max.KappaSD.fit"
## [19] "min.loss.errorSD.fit" "min.Accuracy.fit"
## [21] "min.Kappa.fit" "min.loss.error.fit"
## [23] "min.loss.error.OOB"
## model_id model_method
## 1 Baseline.mybaseln_classfr mybaseln_classfr
## 2 MFO.myMFO_classfr myMFO_classfr
## 3 Random.myrandom_classfr myrandom_classfr
## 4 Max.cor.Y.cv.0.rpart rpart
## 5 Max.cor.Y.cv.G.rpart rpart
## 6 Interact.High.cor.y.rpart rpart
## 7 Low.cor.X.rpart rpart
## 8 All.X.lser.no.cp.opt.rpart rpart
## 9 All.X.lser.no.cp.4015.rpart rpart
## 10 All.X.lser.ys.cp.opt.rpart rpart
## 11 All.X.lser.ys.cp.4015.rpart rpart
## 12 All.X.lser.no.cp.opt.rf rf
## 13 All.X.lser.ys.cp.opt.rf rf
## feats
## 1 bucket2008.fctr, .rnorm
## 2 .rnorm
## 3 .rnorm
## 4 bucket2008
## 5 bucket2008
## 6 bucket2008, reimbursement2008
## 7 bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age
## 8 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 9 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 10 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 11 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 12 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## 13 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 0 0.626 0.006
## 2 0 0.515 0.038
## 3 0 0.375 0.035
## 4 0 4.811 3.430
## 5 3 20.801 3.333
## 6 3 31.301 5.050
## 7 3 101.182 13.937
## 8 3 107.610 15.544
## 9 1 73.967 15.630
## 10 3 108.644 15.505
## 11 1 75.838 16.151
## 12 3 3479.970 449.343
## 13 3 4381.707 1409.455
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.6829729 0.6812294 0.6847125
## 2 0.6712663 0.6695064 0.6730227
## 3 0.4967267 0.4948556 0.4985980
## 4 0.6712663 0.6695064 0.6730227
## 5 0.7031401 0.7014279 0.7048479
## 6 0.7029800 0.7018289 0.7052475
## 7 0.7086058 0.7063706 0.7097740
## 8 0.7086604 0.7064253 0.7098285
## 9 0.7117062 0.7188342 0.7221934
## 10 0.7035404 0.7018289 0.7052475
## 11 0.7205162 0.7188342 0.7221934
## 12 0.7139405 0.7182509 0.7216123
## 13 1.0000000 0.9999866 1.0000000
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.6838135 0.6816786
## 2 NA 0.6712700 0.6691135
## 3 NA 0.4948800 0.4925879
## 4 NA 0.6712700 0.6691135
## 5 0.3440289 0.7040207 0.7019245
## 6 0.3368565 0.7039770 0.7018807
## 7 0.3161651 0.7077434 0.7056548
## 8 0.3140169 0.7074814 0.7053922
## 9 0.3328722 0.7126669 0.7105887
## 10 NA 0.7039770 0.7018807
## 11 NA 0.7126669 0.7105887
## 12 0.3281909 0.7144900 0.7124157
## 13 NA 0.6939553 0.6918392
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.6859425 NA 0 NA
## 2 0.6734210 NA 0 NA
## 3 0.4971722 NA 0 NA
## 4 0.6734210 NA 0 NA
## 5 0.7061105 NA 0 0.0016657885
## 6 0.7060669 NA 0 0.0016900317
## 7 0.7098254 NA 0 0.0009742811
## 8 0.7095639 NA 0 0.0010140545
## 9 0.7147384 NA 0 0.0021496212
## 10 0.7060669 NA 0 NA
## 11 0.7147384 NA 0 NA
## 12 0.7165575 NA 0 NA
## 13 0.6960652 NA 0 NA
## max.KappaSD.fit min.loss.errorSD.fit min.Accuracy.fit min.Kappa.fit
## 1 NA NA NA NA
## 2 NA NA NA NA
## 3 NA NA NA NA
## 4 NA NA NA NA
## 5 0.004508795 NA NA NA
## 6 0.006361514 NA NA NA
## 7 0.004868262 NA NA NA
## 8 0.004809558 NA NA NA
## 9 0.007338010 NA NA NA
## 10 NA 0.016425281 NA NA
## 11 NA 0.005993538 NA NA
## 12 NA NA NA NA
## 13 NA NA 0.6943265 0.3128527
## min.loss.error.fit min.loss.error.OOB
## 1 0.7402976 0.7386055
## 2 1.0443336 1.0443008
## 3 1.1765119 1.1735298
## 4 1.0443336 1.0443008
## 5 0.7424628 0.7406251
## 6 0.7455013 0.7441403
## 7 0.7837833 0.7846967
## 8 0.7861887 0.7875787
## 9 0.7447408 0.7578902
## 10 0.7455013 0.7441403
## 11 0.7447408 0.7578902
## 12 0.7583760 0.7668584
## 13 0.0000000 0.7655702plt_models_df <- glb_models_df[, -grep("SD|Upper|Lower", names(glb_models_df))]
for (var in grep("^min.", names(plt_models_df), value=TRUE)) {
plt_models_df[, sub("min.", "inv.", var)] <-
#ifelse(all(is.na(tmp <- plt_models_df[, var])), NA, 1.0 / tmp)
1.0 / plt_models_df[, var]
plt_models_df <- plt_models_df[ , -grep(var, names(plt_models_df))]
}
print(plt_models_df)## model_id model_method
## 1 Baseline.mybaseln_classfr mybaseln_classfr
## 2 MFO.myMFO_classfr myMFO_classfr
## 3 Random.myrandom_classfr myrandom_classfr
## 4 Max.cor.Y.cv.0.rpart rpart
## 5 Max.cor.Y.cv.G.rpart rpart
## 6 Interact.High.cor.y.rpart rpart
## 7 Low.cor.X.rpart rpart
## 8 All.X.lser.no.cp.opt.rpart rpart
## 9 All.X.lser.no.cp.4015.rpart rpart
## 10 All.X.lser.ys.cp.opt.rpart rpart
## 11 All.X.lser.ys.cp.4015.rpart rpart
## 12 All.X.lser.no.cp.opt.rf rf
## 13 All.X.lser.ys.cp.opt.rf rf
## feats
## 1 bucket2008.fctr, .rnorm
## 2 .rnorm
## 3 .rnorm
## 4 bucket2008
## 5 bucket2008
## 6 bucket2008, reimbursement2008
## 7 bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age
## 8 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 9 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 10 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 11 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 12 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## 13 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## max.nTuningRuns max.Accuracy.fit max.Kappa.fit max.Accuracy.OOB
## 1 0 0.6829729 NA 0.6838135
## 2 0 0.6712663 NA 0.6712700
## 3 0 0.4967267 NA 0.4948800
## 4 0 0.6712663 NA 0.6712700
## 5 3 0.7031401 0.3440289 0.7040207
## 6 3 0.7029800 0.3368565 0.7039770
## 7 3 0.7086058 0.3161651 0.7077434
## 8 3 0.7086604 0.3140169 0.7074814
## 9 1 0.7117062 0.3328722 0.7126669
## 10 3 0.7035404 NA 0.7039770
## 11 1 0.7205162 NA 0.7126669
## 12 3 0.7139405 0.3281909 0.7144900
## 13 3 1.0000000 NA 0.6939553
## max.Kappa.OOB inv.elapsedtime.everything inv.elapsedtime.final
## 1 NA 1.5974440895 1.666667e+02
## 2 NA 1.9417475728 2.631579e+01
## 3 NA 2.6666666667 2.857143e+01
## 4 NA 0.2078569944 2.915452e-01
## 5 NA 0.0480746118 3.000300e-01
## 6 NA 0.0319478611 1.980198e-01
## 7 NA 0.0098831808 7.175145e-02
## 8 NA 0.0092928167 6.433350e-02
## 9 NA 0.0135195425 6.397953e-02
## 10 NA 0.0092043739 6.449532e-02
## 11 NA 0.0131860017 6.191567e-02
## 12 NA 0.0002873588 2.225471e-03
## 13 NA 0.0002282216 7.094941e-04
## inv.SSE.fit inv.Accuracy.fit inv.Kappa.fit inv.loss.error.fit
## 1 Inf NA NA 1.3508081
## 2 Inf NA NA 0.9575485
## 3 Inf NA NA 0.8499702
## 4 Inf NA NA 0.9575485
## 5 Inf NA NA 1.3468689
## 6 Inf NA NA 1.3413792
## 7 Inf NA NA 1.2758629
## 8 Inf NA NA 1.2719593
## 9 Inf NA NA 1.3427491
## 10 Inf NA NA 1.3413792
## 11 Inf NA NA 1.3427491
## 12 Inf NA NA 1.3186071
## 13 Inf 1.440245 3.196392 Inf
## inv.loss.error.OOB
## 1 1.3539028
## 2 0.9575785
## 3 0.8521301
## 4 0.9575785
## 5 1.3502108
## 6 1.3438325
## 7 1.2743778
## 8 1.2697143
## 9 1.3194523
## 10 1.3438325
## 11 1.3194523
## 12 1.3040216
## 13 1.3062159print(myplot_radar(radar_inp_df=plt_models_df))## Warning in myplot_radar(radar_inp_df = plt_models_df): Not plotting
## columns with all NAs: max.Kappa.OOB## Warning in myplot_radar(radar_inp_df = plt_models_df): Not plotting
## columns with all Infs: inv.SSE.fit,inv.loss.error.fit## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 13. Consider specifying shapes manually. if you must have them.## Warning: Removed 83 rows containing missing values (geom_point).## Warning: Removed 31 rows containing missing values (geom_text).## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 13. Consider specifying shapes manually. if you must have them.
print(myplot_radar(radar_inp_df=subset(plt_models_df,
!(model_id %in% grep("random|MFO", plt_models_df$model_id, value=TRUE)))))## Warning in myplot_radar(radar_inp_df = subset(plt_models_df, !(model_id
## %in% : Not plotting columns with all NAs: max.Kappa.OOB## Warning in myplot_radar(radar_inp_df = subset(plt_models_df, !(model_id
## %in% : Not plotting columns with all Infs: inv.SSE.fit,inv.loss.error.fit## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 11. Consider specifying shapes manually. if you must have them.## Warning: Removed 63 rows containing missing values (geom_point).## Warning: Removed 25 rows containing missing values (geom_text).## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 11. Consider specifying shapes manually. if you must have them.
# Compute CI for <metric>SD
glb_models_df <- mutate(glb_models_df,
max.df = ifelse(max.nTuningRuns > 1, max.nTuningRuns - 1, NA),
min.sd2ci.scaler = ifelse(is.na(max.df), NA, qt(0.975, max.df)))
for (var in grep("SD", names(glb_models_df), value=TRUE)) {
# Does CI alredy exist ?
var_components <- unlist(strsplit(var, "SD"))
varActul <- paste0(var_components[1], var_components[2])
varUpper <- paste0(var_components[1], "Upper", var_components[2])
varLower <- paste0(var_components[1], "Lower", var_components[2])
if (varUpper %in% names(glb_models_df)) {
warning(varUpper, " already exists in glb_models_df")
# Assuming Lower also exists
next
}
print(sprintf("var:%s", var))
# CI is dependent on sample size in t distribution; df=n-1
glb_models_df[, varUpper] <- glb_models_df[, varActul] +
glb_models_df[, "min.sd2ci.scaler"] * glb_models_df[, var]
glb_models_df[, varLower] <- glb_models_df[, varActul] -
glb_models_df[, "min.sd2ci.scaler"] * glb_models_df[, var]
}## Warning: max.AccuracyUpper.fit already exists in glb_models_df## [1] "var:max.KappaSD.fit"
## [1] "var:min.loss.errorSD.fit"# Plot metrics with CI
plt_models_df <- glb_models_df[, "model_id", FALSE]
pltCI_models_df <- glb_models_df[, "model_id", FALSE]
for (var in grep("Upper", names(glb_models_df), value=TRUE)) {
var_components <- unlist(strsplit(var, "Upper"))
col_name <- unlist(paste(var_components, collapse=""))
plt_models_df[, col_name] <- glb_models_df[, col_name]
for (name in paste0(var_components[1], c("Upper", "Lower"), var_components[2]))
pltCI_models_df[, name] <- glb_models_df[, name]
}
build_statsCI_data <- function(plt_models_df) {
mltd_models_df <- melt(plt_models_df, id.vars="model_id")
mltd_models_df$data <- sapply(1:nrow(mltd_models_df),
function(row_ix) tail(unlist(strsplit(as.character(
mltd_models_df[row_ix, "variable"]), "[.]")), 1))
mltd_models_df$label <- sapply(1:nrow(mltd_models_df),
function(row_ix) head(unlist(strsplit(as.character(
mltd_models_df[row_ix, "variable"]), paste0(".", mltd_models_df[row_ix, "data"]))), 1))
#print(mltd_models_df)
return(mltd_models_df)
}
mltd_models_df <- build_statsCI_data(plt_models_df)
mltdCI_models_df <- melt(pltCI_models_df, id.vars="model_id")
for (row_ix in 1:nrow(mltdCI_models_df)) {
for (type in c("Upper", "Lower")) {
if (length(var_components <- unlist(strsplit(
as.character(mltdCI_models_df[row_ix, "variable"]), type))) > 1) {
#print(sprintf("row_ix:%d; type:%s; ", row_ix, type))
mltdCI_models_df[row_ix, "label"] <- var_components[1]
mltdCI_models_df[row_ix, "data"] <- unlist(strsplit(var_components[2], "[.]"))[2]
mltdCI_models_df[row_ix, "type"] <- type
break
}
}
}
#print(mltdCI_models_df)
# castCI_models_df <- dcast(mltdCI_models_df, value ~ type, fun.aggregate=sum)
# print(castCI_models_df)
wideCI_models_df <- reshape(subset(mltdCI_models_df, select=-variable),
timevar="type",
idvar=setdiff(names(mltdCI_models_df), c("type", "value", "variable")),
direction="wide")
#print(wideCI_models_df)
mrgdCI_models_df <- merge(wideCI_models_df, mltd_models_df, all.x=TRUE)
#print(mrgdCI_models_df)
# Merge stats back in if CIs don't exist
goback_vars <- c()
for (var in unique(mltd_models_df$label)) {
for (type in unique(mltd_models_df$data)) {
var_type <- paste0(var, ".", type)
# if this data is already present, next
if (var_type %in% unique(paste(mltd_models_df$label, mltd_models_df$data, sep=".")))
next
#print(sprintf("var_type:%s", var_type))
goback_vars <- c(goback_vars, var_type)
}
}
mltd_goback_df <- build_statsCI_data(glb_models_df[, c("model_id", goback_vars)])
mltd_models_df <- rbind(mltd_models_df, mltd_goback_df)
mltd_models_df <- merge(mltd_models_df, glb_models_df[, c("model_id", "model_method")], all.x=TRUE)
# print(myplot_bar(mltd_models_df, "model_id", "value", colorcol_name="data") +
# geom_errorbar(data=mrgdCI_models_df,
# mapping=aes(x=model_id, ymax=value.Upper, ymin=value.Lower), width=0.5) +
# facet_grid(label ~ data, scales="free") +
# theme(axis.text.x = element_text(angle = 45,vjust = 1)))
# mltd_models_df <- orderBy(~ value +variable +data +label + model_method + model_id,
# mltd_models_df)
print(myplot_bar(mltd_models_df, "model_id", "value", colorcol_name="model_method") +
geom_errorbar(data=mrgdCI_models_df,
mapping=aes(x=model_id, ymax=value.Upper, ymin=value.Lower), width=0.5) +
facet_grid(label ~ data, scales="free") +
theme(axis.text.x = element_text(angle = 90,vjust = 0.5)))## Warning: Removed 7 rows containing missing values (position_stack).## Warning: Removed 13 rows containing missing values (position_stack).## Warning in min(x): no non-missing arguments to min; returning Inf## Warning in max(x): no non-missing arguments to max; returning -Inf## Warning: position_stack requires constant width: output may be incorrect
if (glb_is_regression) {
print(orderBy(~ -R.sq.OOB -Adj.R.sq.fit, glb_models_df))
stop("glb_sel_mdl not selected")
print(myplot_scatter(plot_models_df, "Adj.R.sq.fit", "R.sq.OOB") +
geom_text(aes(label=feats.label), data=plot_models_df, color="NavyBlue",
size=3.5, angle=45))
}
if (glb_is_classification) {
print(tmp_models_df <- orderBy(glb_model_sel_frmla, glb_models_df))
print("Metrics used for model selection:"); print(glb_model_sel_frmla)
print(sprintf("Best model id: %s", tmp_models_df[1, "model_id"]))
if (is.null(glb_sel_mdl_id))
{ glb_sel_mdl_id <- tmp_models_df[1, "model_id"] } else
print(sprintf("User specified selection: %s", glb_sel_mdl_id))
myprint_mdl(glb_sel_mdl <- glb_models_lst[[glb_sel_mdl_id]])
} ## model_id model_method
## 1 Baseline.mybaseln_classfr mybaseln_classfr
## 5 Max.cor.Y.cv.G.rpart rpart
## 6 Interact.High.cor.y.rpart rpart
## 10 All.X.lser.ys.cp.opt.rpart rpart
## 9 All.X.lser.no.cp.4015.rpart rpart
## 11 All.X.lser.ys.cp.4015.rpart rpart
## 13 All.X.lser.ys.cp.opt.rf rf
## 12 All.X.lser.no.cp.opt.rf rf
## 7 Low.cor.X.rpart rpart
## 8 All.X.lser.no.cp.opt.rpart rpart
## 2 MFO.myMFO_classfr myMFO_classfr
## 4 Max.cor.Y.cv.0.rpart rpart
## 3 Random.myrandom_classfr myrandom_classfr
## feats
## 1 bucket2008.fctr, .rnorm
## 5 bucket2008
## 6 bucket2008, reimbursement2008
## 10 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 9 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 11 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 13 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## 12 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008, .rnorm
## 7 bucket2008, ihd, diabetes, kidney, heart.failure, copd, depression, alzheimers, arthritis, cancer, osteoporosis, stroke, age
## 8 age, alzheimers, arthritis, cancer, copd, depression, diabetes, heart.failure, ihd, kidney, osteoporosis, stroke, reimbursement2008, bucket2008
## 2 .rnorm
## 4 bucket2008
## 3 .rnorm
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 0 0.626 0.006
## 5 3 20.801 3.333
## 6 3 31.301 5.050
## 10 3 108.644 15.505
## 9 1 73.967 15.630
## 11 1 75.838 16.151
## 13 3 4381.707 1409.455
## 12 3 3479.970 449.343
## 7 3 101.182 13.937
## 8 3 107.610 15.544
## 2 0 0.515 0.038
## 4 0 4.811 3.430
## 3 0 0.375 0.035
## max.Accuracy.fit max.AccuracyLower.fit max.AccuracyUpper.fit
## 1 0.6829729 0.6812294 0.6847125
## 5 0.7031401 0.7014279 0.7048479
## 6 0.7029800 0.7018289 0.7052475
## 10 0.7035404 0.7018289 0.7052475
## 9 0.7117062 0.7188342 0.7221934
## 11 0.7205162 0.7188342 0.7221934
## 13 1.0000000 0.9999866 1.0000000
## 12 0.7139405 0.7182509 0.7216123
## 7 0.7086058 0.7063706 0.7097740
## 8 0.7086604 0.7064253 0.7098285
## 2 0.6712663 0.6695064 0.6730227
## 4 0.6712663 0.6695064 0.6730227
## 3 0.4967267 0.4948556 0.4985980
## max.Kappa.fit max.Accuracy.OOB max.AccuracyLower.OOB
## 1 NA 0.6838135 0.6816786
## 5 0.3440289 0.7040207 0.7019245
## 6 0.3368565 0.7039770 0.7018807
## 10 NA 0.7039770 0.7018807
## 9 0.3328722 0.7126669 0.7105887
## 11 NA 0.7126669 0.7105887
## 13 NA 0.6939553 0.6918392
## 12 0.3281909 0.7144900 0.7124157
## 7 0.3161651 0.7077434 0.7056548
## 8 0.3140169 0.7074814 0.7053922
## 2 NA 0.6712700 0.6691135
## 4 NA 0.6712700 0.6691135
## 3 NA 0.4948800 0.4925879
## max.AccuracyUpper.OOB max.Kappa.OOB min.SSE.fit max.AccuracySD.fit
## 1 0.6859425 NA 0 NA
## 5 0.7061105 NA 0 0.0016657885
## 6 0.7060669 NA 0 0.0016900317
## 10 0.7060669 NA 0 NA
## 9 0.7147384 NA 0 0.0021496212
## 11 0.7147384 NA 0 NA
## 13 0.6960652 NA 0 NA
## 12 0.7165575 NA 0 NA
## 7 0.7098254 NA 0 0.0009742811
## 8 0.7095639 NA 0 0.0010140545
## 2 0.6734210 NA 0 NA
## 4 0.6734210 NA 0 NA
## 3 0.4971722 NA 0 NA
## max.KappaSD.fit min.loss.errorSD.fit min.Accuracy.fit min.Kappa.fit
## 1 NA NA NA NA
## 5 0.004508795 NA NA NA
## 6 0.006361514 NA NA NA
## 10 NA 0.016425281 NA NA
## 9 0.007338010 NA NA NA
## 11 NA 0.005993538 NA NA
## 13 NA NA 0.6943265 0.3128527
## 12 NA NA NA NA
## 7 0.004868262 NA NA NA
## 8 0.004809558 NA NA NA
## 2 NA NA NA NA
## 4 NA NA NA NA
## 3 NA NA NA NA
## min.loss.error.fit min.loss.error.OOB max.df min.sd2ci.scaler
## 1 0.7402976 0.7386055 NA NA
## 5 0.7424628 0.7406251 2 4.302653
## 6 0.7455013 0.7441403 2 4.302653
## 10 0.7455013 0.7441403 2 4.302653
## 9 0.7447408 0.7578902 NA NA
## 11 0.7447408 0.7578902 NA NA
## 13 0.0000000 0.7655702 2 4.302653
## 12 0.7583760 0.7668584 2 4.302653
## 7 0.7837833 0.7846967 2 4.302653
## 8 0.7861887 0.7875787 2 4.302653
## 2 1.0443336 1.0443008 NA NA
## 4 1.0443336 1.0443008 NA NA
## 3 1.1765119 1.1735298 NA NA
## max.KappaUpper.fit max.KappaLower.fit min.loss.errorUpper.fit
## 1 NA NA NA
## 5 0.3634286 0.3246291 NA
## 6 0.3642279 0.3094851 NA
## 10 NA NA 0.8161736
## 9 NA NA NA
## 11 NA NA NA
## 13 NA NA NA
## 12 NA NA NA
## 7 0.3371115 0.2952186 NA
## 8 0.3347108 0.2933231 NA
## 2 NA NA NA
## 4 NA NA NA
## 3 NA NA NA
## min.loss.errorLower.fit
## 1 NA
## 5 NA
## 6 NA
## 10 0.674829
## 9 NA
## 11 NA
## 13 NA
## 12 NA
## 7 NA
## 8 NA
## 2 NA
## 4 NA
## 3 NA
## [1] "Metrics used for model selection:"
## ~+min.loss.error.OOB - max.Accuracy.OOB - max.Kappa.OOB
## [1] "Best model id: Baseline.mybaseln_classfr"
## [1] "User specified selection: All.X.lser.ys.cp.4015.rpart"## Warning: labs do not fit even at cex 0.15, there may be some overplotting
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 274803
##
## CP nsplit rel error
## 1 4.908841e-02 0 1.0000000
## 2 1.395884e-02 2 0.9018232
## 3 4.350377e-03 3 0.8878643
## 4 3.597640e-03 4 0.8835140
## 5 9.879673e-04 5 0.8799163
## 6 7.859460e-04 10 0.8748685
## 7 6.789392e-04 11 0.8740826
## 8 4.649258e-04 14 0.8720458
## 9 4.372516e-04 15 0.8715809
## 10 2.988809e-04 19 0.8698319
## 11 2.767415e-04 20 0.8695330
## 12 2.435326e-04 21 0.8692562
## 13 2.036818e-04 22 0.8690127
## 14 1.881842e-04 28 0.8677729
## 15 1.826494e-04 29 0.8675847
## 16 1.605101e-04 31 0.8672194
## 17 1.439056e-04 33 0.8668984
## 18 1.411382e-04 37 0.8663228
## 19 1.217663e-04 42 0.8655922
## 20 1.162314e-04 45 0.8652269
## 21 1.129105e-04 47 0.8649944
## 22 1.051618e-04 52 0.8644299
## 23 9.962695e-05 57 0.8638764
## 24 9.409212e-05 68 0.8627473
## 25 8.855729e-05 74 0.8621827
## 26 8.302246e-05 83 0.8613746
## 27 7.748763e-05 85 0.8612086
## 28 7.379774e-05 97 0.8602455
## 29 7.195280e-05 101 0.8599134
## 30 6.918538e-05 114 0.8588729
## 31 6.641797e-05 122 0.8583194
## 32 6.272808e-05 154 0.8560612
## 33 6.167383e-05 158 0.8557955
## 34 6.088314e-05 166 0.8552752
## 35 5.811572e-05 179 0.8544340
## 36 5.534831e-05 183 0.8542015
## 37 5.313437e-05 228 0.8516001
## 38 5.258089e-05 233 0.8513344
## 39 5.165842e-05 237 0.8511241
## 40 5.000000e-05 254 0.8501832
##
## Variable importance
## reimbursement2008 bucket2008 diabetes ihd
## 31 20 13 13
## heart.failure kidney arthritis
## 11 9 1
##
## Node number 1: 274803 observations, complexity param=0.04908841
## predicted class=B1 expected loss=0.3287337 P(node) =1
## class counts: 184466 52259 24586 11906 1586
## probabilities: 0.671 0.190 0.089 0.043 0.006
## left son=2 (165987 obs) right son=3 (108816 obs)
## Primary splits:
## reimbursement2008 < 1565 to the left, improve=24395.14, (0 missing)
## bucket2008 < 1.5 to the left, improve=20624.70, (0 missing)
## ihd < 0.5 to the left, improve=16291.74, (0 missing)
## diabetes < 0.5 to the left, improve=16041.26, (0 missing)
## heart.failure < 0.5 to the left, improve=12498.16, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.861, adj=0.650, (0 split)
## ihd < 0.5 to the left, agree=0.792, adj=0.474, (0 split)
## diabetes < 0.5 to the left, agree=0.785, adj=0.456, (0 split)
## heart.failure < 0.5 to the left, agree=0.762, adj=0.399, (0 split)
## kidney < 0.5 to the left, agree=0.731, adj=0.321, (0 split)
##
## Node number 2: 165987 observations
## predicted class=B1 expected loss=0.1261424 P(node) =0.6040218
## class counts: 145049 12284 6102 2315 237
## probabilities: 0.874 0.074 0.037 0.014 0.001
##
## Node number 3: 108816 observations, complexity param=0.04908841
## predicted class=B2 expected loss=0.6326367 P(node) =0.3959782
## class counts: 39417 39975 18484 9591 1349
## probabilities: 0.362 0.367 0.170 0.088 0.012
## left son=6 (39298 obs) right son=7 (69518 obs)
## Primary splits:
## reimbursement2008 < 3065 to the left, improve=2010.3080, (0 missing)
## bucket2008 < 1.5 to the left, improve=1980.9770, (0 missing)
## kidney < 0.5 to the left, improve=1416.9220, (0 missing)
## diabetes < 0.5 to the left, improve=1236.1460, (0 missing)
## heart.failure < 0.5 to the left, improve= 976.9427, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.989, adj=0.969, (0 split)
## ihd < 0.5 to the left, agree=0.659, adj=0.056, (0 split)
## diabetes < 0.5 to the left, agree=0.641, adj=0.006, (0 split)
##
## Node number 6: 39298 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.4797445 P(node) =0.1430043
## class counts: 20445 12134 4756 1782 181
## probabilities: 0.520 0.309 0.121 0.045 0.005
## left son=12 (20077 obs) right son=13 (19221 obs)
## Primary splits:
## reimbursement2008 < 2175 to the left, improve=192.7592, (0 missing)
## diabetes < 0.5 to the left, improve=155.3521, (0 missing)
## ihd < 0.5 to the left, improve=114.8541, (0 missing)
## arthritis < 0.5 to the left, improve=114.6837, (0 missing)
## kidney < 0.5 to the left, improve=108.9096, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.542, adj=0.063, (0 split)
## arthritis < 0.5 to the left, agree=0.539, adj=0.058, (0 split)
## ihd < 0.5 to the left, agree=0.534, adj=0.048, (0 split)
## kidney < 0.5 to the left, agree=0.532, adj=0.044, (0 split)
## diabetes < 0.5 to the left, agree=0.532, adj=0.043, (0 split)
##
## Node number 7: 69518 observations, complexity param=0.01395884
## predicted class=B2 expected loss=0.5995138 P(node) =0.2529739
## class counts: 18972 27841 13728 7809 1168
## probabilities: 0.273 0.400 0.197 0.112 0.017
## left son=14 (15717 obs) right son=15 (53801 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=646.4740, (0 missing)
## kidney < 0.5 to the left, improve=604.0313, (0 missing)
## arthritis < 0.5 to the left, improve=501.1263, (0 missing)
## ihd < 0.5 to the left, improve=427.9009, (0 missing)
## heart.failure < 0.5 to the left, improve=380.0080, (0 missing)
##
## Node number 12: 20077 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.4247148 P(node) =0.07305961
## class counts: 11550 5416 2200 834 77
## probabilities: 0.575 0.270 0.110 0.042 0.004
## left son=24 (8826 obs) right son=25 (11251 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=62.34344, (0 missing)
## kidney < 0.5 to the left, improve=42.15624, (0 missing)
## ihd < 0.5 to the left, improve=40.01287, (0 missing)
## heart.failure < 0.5 to the left, improve=36.00697, (0 missing)
## arthritis < 0.5 to the left, improve=33.77686, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.588, adj=0.062, (0 split)
##
## Node number 13: 19221 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.5372249 P(node) =0.06994465
## class counts: 8895 6718 2556 948 104
## probabilities: 0.463 0.350 0.133 0.049 0.005
## left son=26 (7137 obs) right son=27 (12084 obs)
## Primary splits:
## diabetes < 0.5 to the left, improve=71.31724, (0 missing)
## arthritis < 0.5 to the left, improve=61.00585, (0 missing)
## ihd < 0.5 to the left, improve=55.20411, (0 missing)
## heart.failure < 0.5 to the left, improve=52.20163, (0 missing)
## kidney < 0.5 to the left, improve=49.73230, (0 missing)
##
## Node number 14: 15717 observations, complexity param=0.004350377
## predicted class=B1 expected loss=0.5704651 P(node) =0.0571937
## class counts: 6751 5490 2365 999 112
## probabilities: 0.430 0.349 0.150 0.064 0.007
## left son=28 (13123 obs) right son=29 (2594 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=130.12270, (0 missing)
## arthritis < 0.5 to the left, improve=125.41530, (0 missing)
## ihd < 0.5 to the left, improve= 80.76118, (0 missing)
## depression < 0.5 to the left, improve= 61.32779, (0 missing)
## osteoporosis < 0.5 to the left, improve= 44.50253, (0 missing)
##
## Node number 15: 53801 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.5845616 P(node) =0.1957802
## class counts: 12221 22351 11363 6810 1056
## probabilities: 0.227 0.415 0.211 0.127 0.020
## left son=30 (25067 obs) right son=31 (28734 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=408.9756, (0 missing)
## reimbursement2008 < 15395 to the left, improve=327.1281, (0 missing)
## bucket2008 < 3.5 to the left, improve=313.8191, (0 missing)
## arthritis < 0.5 to the left, improve=266.5595, (0 missing)
## heart.failure < 0.5 to the left, improve=209.4718, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8365 to the left, agree=0.666, adj=0.282, (0 split)
## bucket2008 < 2.5 to the left, agree=0.664, adj=0.279, (0 split)
## heart.failure < 0.5 to the left, agree=0.628, adj=0.201, (0 split)
## copd < 0.5 to the left, agree=0.595, adj=0.132, (0 split)
## ihd < 0.5 to the left, agree=0.575, adj=0.089, (0 split)
##
## Node number 24: 8826 observations
## predicted class=B1 expected loss=0.3716293 P(node) =0.03211755
## class counts: 5546 2137 805 312 26
## probabilities: 0.628 0.242 0.091 0.035 0.003
##
## Node number 25: 11251 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.4663585 P(node) =0.04094206
## class counts: 6004 3279 1395 522 51
## probabilities: 0.534 0.291 0.124 0.046 0.005
## left son=50 (9007 obs) right son=51 (2244 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=18.86048, (0 missing)
## heart.failure < 0.5 to the left, improve=17.29926, (0 missing)
## arthritis < 0.5 to the left, improve=16.91283, (0 missing)
## reimbursement2008 < 1875 to the left, improve=16.48954, (0 missing)
## cancer < 0.5 to the left, improve=14.98495, (0 missing)
##
## Node number 26: 7137 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.470786 P(node) =0.02597133
## class counts: 3777 2233 794 300 33
## probabilities: 0.529 0.313 0.111 0.042 0.005
## left son=52 (5554 obs) right son=53 (1583 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=24.840370, (0 missing)
## depression < 0.5 to the left, improve=16.217060, (0 missing)
## ihd < 0.5 to the left, improve=13.895180, (0 missing)
## copd < 0.5 to the left, improve=12.688930, (0 missing)
## kidney < 0.5 to the left, improve= 9.728645, (0 missing)
##
## Node number 27: 12084 observations, complexity param=0.0006789392
## predicted class=B1 expected loss=0.5764647 P(node) =0.04397332
## class counts: 5118 4485 1762 648 71
## probabilities: 0.424 0.371 0.146 0.054 0.006
## left son=54 (8413 obs) right son=55 (3671 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=27.83165, (0 missing)
## heart.failure < 0.5 to the left, improve=26.70933, (0 missing)
## ihd < 0.5 to the left, improve=24.37311, (0 missing)
## kidney < 0.5 to the left, improve=22.60183, (0 missing)
## reimbursement2008 < 2655 to the left, improve=21.75660, (0 missing)
##
## Node number 28: 13123 observations, complexity param=0.00359764
## predicted class=B1 expected loss=0.5360055 P(node) =0.04775421
## class counts: 6089 4435 1751 763 85
## probabilities: 0.464 0.338 0.133 0.058 0.006
## left son=56 (9625 obs) right son=57 (3498 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=126.28480, (0 missing)
## ihd < 0.5 to the left, improve= 70.76778, (0 missing)
## depression < 0.5 to the left, improve= 68.94332, (0 missing)
## osteoporosis < 0.5 to the left, improve= 46.31934, (0 missing)
## heart.failure < 0.5 to the left, improve= 30.26771, (0 missing)
##
## Node number 29: 2594 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.5932922 P(node) =0.009439489
## class counts: 662 1055 614 236 27
## probabilities: 0.255 0.407 0.237 0.091 0.010
## left son=58 (1000 obs) right son=59 (1594 obs)
## Primary splits:
## reimbursement2008 < 5770 to the left, improve=8.464458, (0 missing)
## arthritis < 0.5 to the left, improve=7.371565, (0 missing)
## ihd < 0.5 to the left, improve=5.410820, (0 missing)
## copd < 0.5 to the left, improve=5.301788, (0 missing)
## heart.failure < 0.5 to the left, improve=3.070575, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.823, adj=0.542, (0 split)
##
## Node number 30: 25067 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.57091 P(node) =0.09121807
## class counts: 7517 10756 4691 1917 186
## probabilities: 0.300 0.429 0.187 0.076 0.007
## left son=60 (15178 obs) right son=61 (9889 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=169.25970, (0 missing)
## cancer < 0.5 to the left, improve= 99.57556, (0 missing)
## ihd < 0.5 to the left, improve= 68.28883, (0 missing)
## depression < 0.5 to the left, improve= 61.94482, (0 missing)
## heart.failure < 0.5 to the left, improve= 42.19646, (0 missing)
## Surrogate splits:
## reimbursement2008 < 66495 to the left, agree=0.606, adj=0.001, (0 split)
##
## Node number 31: 28734 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.5964711 P(node) =0.1045622
## class counts: 4704 11595 6672 4893 870
## probabilities: 0.164 0.404 0.232 0.170 0.030
## left son=62 (16249 obs) right son=63 (12485 obs)
## Primary splits:
## reimbursement2008 < 15395 to the left, improve=177.49270, (0 missing)
## bucket2008 < 3.5 to the left, improve=170.28940, (0 missing)
## arthritis < 0.5 to the left, improve=101.31920, (0 missing)
## heart.failure < 0.5 to the left, improve= 62.82321, (0 missing)
## ihd < 0.5 to the left, improve= 55.35075, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.924, adj=0.826, (0 split)
## copd < 0.5 to the left, agree=0.609, adj=0.101, (0 split)
## stroke < 0.5 to the left, agree=0.605, adj=0.091, (0 split)
## cancer < 0.5 to the left, agree=0.580, adj=0.033, (0 split)
## alzheimers < 0.5 to the left, agree=0.569, adj=0.008, (0 split)
##
## Node number 50: 9007 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.4490951 P(node) =0.03277621
## class counts: 4962 2540 1087 378 40
## probabilities: 0.551 0.282 0.121 0.042 0.004
## left son=100 (4935 obs) right son=101 (4072 obs)
## Primary splits:
## reimbursement2008 < 1875 to the left, improve=14.670650, (0 missing)
## cancer < 0.5 to the left, improve=12.077140, (0 missing)
## arthritis < 0.5 to the left, improve= 9.470091, (0 missing)
## heart.failure < 0.5 to the left, improve= 7.308909, (0 missing)
## depression < 0.5 to the left, improve= 6.801973, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.549, adj=0.003, (0 split)
## age < 29.5 to the right, agree=0.548, adj=0.001, (0 split)
##
## Node number 51: 2244 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5356506 P(node) =0.00816585
## class counts: 1042 739 308 144 11
## probabilities: 0.464 0.329 0.137 0.064 0.005
## left son=102 (992 obs) right son=103 (1252 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=7.795458, (0 missing)
## arthritis < 0.5 to the left, improve=7.027320, (0 missing)
## ihd < 0.5 to the left, improve=4.964222, (0 missing)
## reimbursement2008 < 1735 to the left, improve=4.132280, (0 missing)
## cancer < 0.5 to the left, improve=3.835396, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.565, adj=0.016, (0 split)
## age < 33.5 to the left, agree=0.559, adj=0.002, (0 split)
##
## Node number 52: 5554 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4449046 P(node) =0.02021084
## class counts: 3083 1647 580 217 27
## probabilities: 0.555 0.297 0.104 0.039 0.005
## left son=104 (2348 obs) right son=105 (3206 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=13.118310, (0 missing)
## depression < 0.5 to the left, improve=12.689550, (0 missing)
## kidney < 0.5 to the left, improve= 9.684755, (0 missing)
## copd < 0.5 to the left, improve= 9.145592, (0 missing)
## heart.failure < 0.5 to the left, improve= 8.228139, (0 missing)
## Surrogate splits:
## age < 28.5 to the left, agree=0.579, adj=0.004, (0 split)
## reimbursement2008 < 2185 to the left, agree=0.578, adj=0.001, (0 split)
##
## Node number 53: 1583 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5615919 P(node) =0.00576049
## class counts: 694 586 214 83 6
## probabilities: 0.438 0.370 0.135 0.052 0.004
## left son=106 (1525 obs) right son=107 (58 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=5.133391, (0 missing)
## reimbursement2008 < 2725 to the left, improve=3.164238, (0 missing)
## cancer < 0.5 to the left, improve=2.451745, (0 missing)
## copd < 0.5 to the left, improve=2.436381, (0 missing)
## depression < 0.5 to the left, improve=1.979459, (0 missing)
##
## Node number 54: 8413 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.5530726 P(node) =0.03061466
## class counts: 3760 2943 1225 438 47
## probabilities: 0.447 0.350 0.146 0.052 0.006
## left son=108 (4375 obs) right son=109 (4038 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=25.12070, (0 missing)
## ihd < 0.5 to the left, improve=19.50225, (0 missing)
## kidney < 0.5 to the left, improve=18.23799, (0 missing)
## depression < 0.5 to the left, improve=14.07225, (0 missing)
## reimbursement2008 < 2615 to the left, improve=12.21338, (0 missing)
## Surrogate splits:
## kidney < 0.5 to the left, agree=0.569, adj=0.103, (0 split)
## copd < 0.5 to the left, agree=0.568, adj=0.100, (0 split)
## alzheimers < 0.5 to the left, agree=0.546, adj=0.054, (0 split)
## ihd < 0.5 to the left, agree=0.544, adj=0.050, (0 split)
## stroke < 0.5 to the left, agree=0.536, adj=0.034, (0 split)
##
## Node number 55: 3671 observations, complexity param=0.0002988809
## predicted class=B2 expected loss=0.579951 P(node) =0.01335866
## class counts: 1358 1542 537 210 24
## probabilities: 0.370 0.420 0.146 0.057 0.007
## left son=110 (2068 obs) right son=111 (1603 obs)
## Primary splits:
## reimbursement2008 < 2665 to the left, improve=10.442080, (0 missing)
## cancer < 0.5 to the left, improve= 4.234333, (0 missing)
## ihd < 0.5 to the left, improve= 4.129116, (0 missing)
## kidney < 0.5 to the left, improve= 3.679214, (0 missing)
## copd < 0.5 to the left, improve= 3.281268, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.638, adj=0.170, (0 split)
## cancer < 0.5 to the left, agree=0.566, adj=0.006, (0 split)
## age < 26.5 to the right, agree=0.564, adj=0.001, (0 split)
## stroke < 0.5 to the left, agree=0.564, adj=0.001, (0 split)
##
## Node number 56: 9625 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.4874805 P(node) =0.03502509
## class counts: 4933 2954 1162 520 56
## probabilities: 0.513 0.307 0.121 0.054 0.006
## left son=112 (3135 obs) right son=113 (6490 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=51.18602, (0 missing)
## depression < 0.5 to the left, improve=46.82343, (0 missing)
## heart.failure < 0.5 to the left, improve=27.25528, (0 missing)
## osteoporosis < 0.5 to the left, improve=25.54800, (0 missing)
## reimbursement2008 < 6615 to the left, improve=12.84564, (0 missing)
##
## Node number 57: 3498 observations, complexity param=0.0004372516
## predicted class=B2 expected loss=0.5766152 P(node) =0.01272912
## class counts: 1156 1481 589 243 29
## probabilities: 0.330 0.423 0.168 0.069 0.008
## left son=114 (2340 obs) right son=115 (1158 obs)
## Primary splits:
## reimbursement2008 < 8525 to the left, improve=12.263650, (0 missing)
## depression < 0.5 to the left, improve=10.454350, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.052395, (0 missing)
## copd < 0.5 to the left, improve= 8.848663, (0 missing)
## ihd < 0.5 to the left, improve= 8.087092, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.979, adj=0.935, (0 split)
## kidney < 0.5 to the left, agree=0.692, adj=0.069, (0 split)
## stroke < 0.5 to the left, agree=0.680, adj=0.033, (0 split)
##
## Node number 58: 1000 observations
## predicted class=B2 expected loss=0.562 P(node) =0.00363897
## class counts: 296 438 191 70 5
## probabilities: 0.296 0.438 0.191 0.070 0.005
##
## Node number 59: 1594 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6129235 P(node) =0.005800519
## class counts: 366 617 423 166 22
## probabilities: 0.230 0.387 0.265 0.104 0.014
## left son=118 (1054 obs) right son=119 (540 obs)
## Primary splits:
## reimbursement2008 < 8645 to the right, improve=7.014383, (0 missing)
## arthritis < 0.5 to the left, improve=5.636989, (0 missing)
## bucket2008 < 2.5 to the right, improve=4.256675, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.245615, (0 missing)
## ihd < 0.5 to the left, improve=2.672736, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.949, adj=0.848, (0 split)
## age < 27.5 to the right, agree=0.662, adj=0.002, (0 split)
##
## Node number 60: 15178 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.6047569 P(node) =0.05523229
## class counts: 5388 5999 2649 1047 95
## probabilities: 0.355 0.395 0.175 0.069 0.006
## left son=120 (12572 obs) right son=121 (2606 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=92.65854, (0 missing)
## ihd < 0.5 to the left, improve=43.72992, (0 missing)
## depression < 0.5 to the left, improve=36.05906, (0 missing)
## heart.failure < 0.5 to the left, improve=30.26654, (0 missing)
## copd < 0.5 to the left, improve=25.73984, (0 missing)
##
## Node number 61: 9889 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5189605 P(node) =0.03598578
## class counts: 2129 4757 2042 870 91
## probabilities: 0.215 0.481 0.206 0.088 0.009
## left son=122 (5134 obs) right son=123 (4755 obs)
## Primary splits:
## depression < 0.5 to the left, improve=18.84327, (0 missing)
## cancer < 0.5 to the left, improve=17.45891, (0 missing)
## ihd < 0.5 to the left, improve=13.35120, (0 missing)
## reimbursement2008 < 9795 to the left, improve=12.33086, (0 missing)
## copd < 0.5 to the left, improve=12.26415, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.564, adj=0.093, (0 split)
## copd < 0.5 to the left, agree=0.546, adj=0.056, (0 split)
## reimbursement2008 < 5815 to the left, agree=0.542, adj=0.048, (0 split)
## age < 64.5 to the right, agree=0.537, adj=0.037, (0 split)
## bucket2008 < 2.5 to the left, agree=0.536, adj=0.036, (0 split)
##
## Node number 62: 16249 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.5619423 P(node) =0.05912963
## class counts: 3113 7118 3819 1946 253
## probabilities: 0.192 0.438 0.235 0.120 0.016
## left son=124 (9424 obs) right son=125 (6825 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=70.60653, (0 missing)
## cancer < 0.5 to the left, improve=30.24922, (0 missing)
## ihd < 0.5 to the left, improve=29.86941, (0 missing)
## reimbursement2008 < 5665 to the left, improve=23.89268, (0 missing)
## bucket2008 < 2.5 to the left, improve=21.55872, (0 missing)
##
## Node number 63: 12485 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.6414097 P(node) =0.04543255
## class counts: 1591 4477 2853 2947 617
## probabilities: 0.127 0.359 0.229 0.236 0.049
## left son=126 (5402 obs) right son=127 (7083 obs)
## Primary splits:
## arthritis < 0.5 to the right, improve=35.40534, (0 missing)
## cancer < 0.5 to the left, improve=26.78171, (0 missing)
## reimbursement2008 < 26625 to the left, improve=24.60405, (0 missing)
## depression < 0.5 to the left, improve=23.29796, (0 missing)
## heart.failure < 0.5 to the left, improve=17.01274, (0 missing)
## Surrogate splits:
## age < 28.5 to the left, agree=0.568, adj=0.002, (0 split)
## reimbursement2008 < 15435 to the left, agree=0.568, adj=0.001, (0 split)
##
## Node number 100: 4935 observations
## predicted class=B1 expected loss=0.4196555 P(node) =0.01795832
## class counts: 2864 1294 550 205 22
## probabilities: 0.580 0.262 0.111 0.042 0.004
##
## Node number 101: 4072 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.4847741 P(node) =0.01481789
## class counts: 2098 1246 537 173 18
## probabilities: 0.515 0.306 0.132 0.042 0.004
## left son=202 (3786 obs) right son=203 (286 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=5.937439, (0 missing)
## arthritis < 0.5 to the left, improve=5.625805, (0 missing)
## copd < 0.5 to the left, improve=3.348444, (0 missing)
## ihd < 0.5 to the left, improve=3.030239, (0 missing)
## heart.failure < 0.5 to the left, improve=2.851779, (0 missing)
##
## Node number 102: 992 observations
## predicted class=B1 expected loss=0.4808468 P(node) =0.003609859
## class counts: 515 292 126 57 2
## probabilities: 0.519 0.294 0.127 0.057 0.002
##
## Node number 103: 1252 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5790735 P(node) =0.004555991
## class counts: 527 447 182 87 9
## probabilities: 0.421 0.357 0.145 0.069 0.007
## left son=206 (904 obs) right son=207 (348 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=7.739842, (0 missing)
## age < 93.5 to the left, improve=3.754099, (0 missing)
## cancer < 0.5 to the left, improve=3.514161, (0 missing)
## reimbursement2008 < 1955 to the left, improve=3.377454, (0 missing)
## ihd < 0.5 to the left, improve=1.751139, (0 missing)
## Surrogate splits:
## age < 30.5 to the right, agree=0.724, adj=0.006, (0 split)
##
## Node number 104: 2348 observations
## predicted class=B1 expected loss=0.3973595 P(node) =0.008544303
## class counts: 1415 632 217 72 12
## probabilities: 0.603 0.269 0.092 0.031 0.005
##
## Node number 105: 3206 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4797255 P(node) =0.01166654
## class counts: 1668 1015 363 145 15
## probabilities: 0.520 0.317 0.113 0.045 0.005
## left son=210 (2325 obs) right son=211 (881 obs)
## Primary splits:
## depression < 0.5 to the left, improve=8.135493, (0 missing)
## kidney < 0.5 to the left, improve=5.219511, (0 missing)
## reimbursement2008 < 2785 to the left, improve=4.205524, (0 missing)
## heart.failure < 0.5 to the left, improve=3.201394, (0 missing)
## copd < 0.5 to the left, improve=3.002159, (0 missing)
## Surrogate splits:
## age < 29.5 to the right, agree=0.726, adj=0.003, (0 split)
##
## Node number 106: 1525 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5534426 P(node) =0.00554943
## class counts: 681 554 202 82 6
## probabilities: 0.447 0.363 0.132 0.054 0.004
## left son=212 (1438 obs) right son=213 (87 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.548424, (0 missing)
## reimbursement2008 < 2715 to the right, improve=2.513748, (0 missing)
## copd < 0.5 to the left, improve=1.973703, (0 missing)
## depression < 0.5 to the left, improve=1.853940, (0 missing)
## kidney < 0.5 to the left, improve=1.632947, (0 missing)
##
## Node number 107: 58 observations
## predicted class=B2 expected loss=0.4482759 P(node) =0.0002110603
## class counts: 13 32 12 1 0
## probabilities: 0.224 0.552 0.207 0.017 0.000
##
## Node number 108: 4375 observations, complexity param=0.0002435326
## predicted class=B1 expected loss=0.5074286 P(node) =0.0159205
## class counts: 2155 1478 555 170 17
## probabilities: 0.493 0.338 0.127 0.039 0.004
## left son=216 (3992 obs) right son=217 (383 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=10.015540, (0 missing)
## ihd < 0.5 to the left, improve= 9.488719, (0 missing)
## depression < 0.5 to the left, improve= 7.316301, (0 missing)
## reimbursement2008 < 2615 to the left, improve= 5.949976, (0 missing)
## copd < 0.5 to the left, improve= 5.117423, (0 missing)
##
## Node number 109: 4038 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.602526 P(node) =0.01469416
## class counts: 1605 1465 670 268 30
## probabilities: 0.397 0.363 0.166 0.066 0.007
## left son=218 (2819 obs) right son=219 (1219 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=10.392200, (0 missing)
## reimbursement2008 < 2455 to the left, improve= 6.028802, (0 missing)
## ihd < 0.5 to the left, improve= 5.795095, (0 missing)
## depression < 0.5 to the left, improve= 5.214940, (0 missing)
## stroke < 0.5 to the left, improve= 3.343262, (0 missing)
##
## Node number 110: 2068 observations, complexity param=0.0002767415
## predicted class=B1 expected loss=0.5918762 P(node) =0.007525391
## class counts: 844 817 280 117 10
## probabilities: 0.408 0.395 0.135 0.057 0.005
## left son=220 (517 obs) right son=221 (1551 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.581883, (0 missing)
## reimbursement2008 < 2305 to the left, improve=3.255344, (0 missing)
## cancer < 0.5 to the left, improve=3.097089, (0 missing)
## age < 54.5 to the left, improve=1.964830, (0 missing)
## kidney < 0.5 to the left, improve=1.730688, (0 missing)
##
## Node number 111: 1603 observations
## predicted class=B2 expected loss=0.547723 P(node) =0.00583327
## class counts: 514 725 257 93 14
## probabilities: 0.321 0.452 0.160 0.058 0.009
##
## Node number 112: 3135 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.3974482 P(node) =0.01140817
## class counts: 1889 825 298 113 10
## probabilities: 0.603 0.263 0.095 0.036 0.003
## left son=224 (2292 obs) right son=225 (843 obs)
## Primary splits:
## depression < 0.5 to the left, improve=19.892930, (0 missing)
## reimbursement2008 < 9505 to the right, improve=15.211730, (0 missing)
## bucket2008 < 2.5 to the right, improve=13.054300, (0 missing)
## osteoporosis < 0.5 to the left, improve=10.317040, (0 missing)
## age < 92.5 to the left, improve= 3.244996, (0 missing)
## Surrogate splits:
## reimbursement2008 < 60755 to the left, agree=0.731, adj=0.001, (0 split)
##
## Node number 113: 6490 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.5309707 P(node) =0.02361692
## class counts: 3044 2129 864 407 46
## probabilities: 0.469 0.328 0.133 0.063 0.007
## left son=226 (4266 obs) right son=227 (2224 obs)
## Primary splits:
## depression < 0.5 to the left, improve=22.130520, (0 missing)
## heart.failure < 0.5 to the left, improve=12.472230, (0 missing)
## osteoporosis < 0.5 to the left, improve=12.135520, (0 missing)
## reimbursement2008 < 6615 to the left, improve=10.028930, (0 missing)
## bucket2008 < 2.5 to the left, improve= 8.000565, (0 missing)
## Surrogate splits:
## age < 34.5 to the right, agree=0.658, adj=0.003, (0 split)
## reimbursement2008 < 115145 to the left, agree=0.658, adj=0.001, (0 split)
##
## Node number 114: 2340 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.542735 P(node) =0.008515191
## class counts: 720 1070 391 144 15
## probabilities: 0.308 0.457 0.167 0.062 0.006
## left son=228 (1359 obs) right son=229 (981 obs)
## Primary splits:
## reimbursement2008 < 4645 to the left, improve=5.782135, (0 missing)
## ihd < 0.5 to the left, improve=5.431632, (0 missing)
## depression < 0.5 to the left, improve=4.505952, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.336155, (0 missing)
## alzheimers < 0.5 to the left, improve=3.247654, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.613, adj=0.076, (0 split)
## copd < 0.5 to the left, agree=0.606, adj=0.059, (0 split)
## kidney < 0.5 to the left, agree=0.586, adj=0.013, (0 split)
## age < 91.5 to the left, agree=0.585, adj=0.011, (0 split)
## stroke < 0.5 to the left, agree=0.585, adj=0.009, (0 split)
##
## Node number 115: 1158 observations, complexity param=0.0004372516
## predicted class=B1 expected loss=0.6234888 P(node) =0.004213928
## class counts: 436 411 198 99 14
## probabilities: 0.377 0.355 0.171 0.085 0.012
## left son=230 (714 obs) right son=231 (444 obs)
## Primary splits:
## copd < 0.5 to the left, improve=13.168040, (0 missing)
## depression < 0.5 to the left, improve= 8.948306, (0 missing)
## kidney < 0.5 to the left, improve= 6.276303, (0 missing)
## ihd < 0.5 to the left, improve= 5.293866, (0 missing)
## reimbursement2008 < 14980 to the left, improve= 4.056180, (0 missing)
## Surrogate splits:
## age < 94.5 to the left, agree=0.626, adj=0.025, (0 split)
## reimbursement2008 < 72745 to the left, agree=0.620, adj=0.009, (0 split)
##
## Node number 118: 1054 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6223909 P(node) =0.003835475
## class counts: 281 398 250 109 16
## probabilities: 0.267 0.378 0.237 0.103 0.015
## left son=236 (745 obs) right son=237 (309 obs)
## Primary splits:
## arthritis < 0.5 to the left, improve=8.492364, (0 missing)
## ihd < 0.5 to the left, improve=3.739184, (0 missing)
## depression < 0.5 to the left, improve=2.714506, (0 missing)
## copd < 0.5 to the left, improve=2.704564, (0 missing)
## reimbursement2008 < 67610 to the left, improve=2.665770, (0 missing)
## Surrogate splits:
## age < 29.5 to the right, agree=0.708, adj=0.003, (0 split)
##
## Node number 119: 540 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.5944444 P(node) =0.001965044
## class counts: 85 219 173 57 6
## probabilities: 0.157 0.406 0.320 0.106 0.011
## left son=238 (243 obs) right son=239 (297 obs)
## Primary splits:
## heart.failure < 0.5 to the right, improve=3.144781, (0 missing)
## reimbursement2008 < 7455 to the left, improve=1.665302, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.352183, (0 missing)
## age < 86.5 to the right, improve=1.232072, (0 missing)
## arthritis < 0.5 to the right, improve=1.028824, (0 missing)
## Surrogate splits:
## copd < 0.5 to the right, agree=0.604, adj=0.119, (0 split)
## kidney < 0.5 to the right, agree=0.585, adj=0.078, (0 split)
## stroke < 0.5 to the right, agree=0.583, adj=0.074, (0 split)
## arthritis < 0.5 to the right, agree=0.576, adj=0.058, (0 split)
## depression < 0.5 to the right, agree=0.572, adj=0.049, (0 split)
##
## Node number 120: 12572 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.613188 P(node) =0.04574914
## class counts: 4844 4863 2000 791 74
## probabilities: 0.385 0.387 0.159 0.063 0.006
## left son=240 (2617 obs) right son=241 (9955 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=36.80981, (0 missing)
## depression < 0.5 to the left, improve=36.47326, (0 missing)
## heart.failure < 0.5 to the left, improve=27.52215, (0 missing)
## copd < 0.5 to the left, improve=21.85222, (0 missing)
## reimbursement2008 < 8955 to the left, improve=19.34797, (0 missing)
##
## Node number 121: 2606 observations
## predicted class=B2 expected loss=0.5640829 P(node) =0.009483157
## class counts: 544 1136 649 256 21
## probabilities: 0.209 0.436 0.249 0.098 0.008
##
## Node number 122: 5134 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5190884 P(node) =0.01868247
## class counts: 1277 2469 936 412 40
## probabilities: 0.249 0.481 0.182 0.080 0.008
## left son=244 (4305 obs) right son=245 (829 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=12.348810, (0 missing)
## reimbursement2008 < 9985 to the left, improve=11.590150, (0 missing)
## bucket2008 < 2.5 to the left, improve= 7.979608, (0 missing)
## ihd < 0.5 to the left, improve= 7.512372, (0 missing)
## copd < 0.5 to the left, improve= 7.186891, (0 missing)
##
## Node number 123: 4755 observations
## predicted class=B2 expected loss=0.5188223 P(node) =0.0173033
## class counts: 852 2288 1106 458 51
## probabilities: 0.179 0.481 0.233 0.096 0.011
##
## Node number 124: 9424 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.5992148 P(node) =0.03429366
## class counts: 2192 3777 2139 1156 160
## probabilities: 0.233 0.401 0.227 0.123 0.017
## left son=248 (7786 obs) right son=249 (1638 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=31.06099, (0 missing)
## ihd < 0.5 to the left, improve=19.93184, (0 missing)
## depression < 0.5 to the left, improve=16.57581, (0 missing)
## reimbursement2008 < 6325 to the left, improve=12.91187, (0 missing)
## bucket2008 < 2.5 to the left, improve=10.82187, (0 missing)
##
## Node number 125: 6825 observations
## predicted class=B2 expected loss=0.5104762 P(node) =0.02483597
## class counts: 921 3341 1680 790 93
## probabilities: 0.135 0.490 0.246 0.116 0.014
##
## Node number 126: 5402 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5960755 P(node) =0.01965772
## class counts: 509 2182 1310 1186 215
## probabilities: 0.094 0.404 0.243 0.220 0.040
## left son=252 (3345 obs) right son=253 (2057 obs)
## Primary splits:
## reimbursement2008 < 34925 to the left, improve=14.212070, (0 missing)
## copd < 0.5 to the left, improve=10.384850, (0 missing)
## depression < 0.5 to the left, improve= 8.104595, (0 missing)
## cancer < 0.5 to the right, improve= 6.743072, (0 missing)
## heart.failure < 0.5 to the left, improve= 6.417519, (0 missing)
## Surrogate splits:
## bucket2008 < 4.5 to the left, agree=0.776, adj=0.413, (0 split)
##
## Node number 127: 7083 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.6759848 P(node) =0.02577483
## class counts: 1082 2295 1543 1761 402
## probabilities: 0.153 0.324 0.218 0.249 0.057
## left son=254 (5298 obs) right son=255 (1785 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=21.09129, (0 missing)
## depression < 0.5 to the left, improve=19.29947, (0 missing)
## reimbursement2008 < 26625 to the left, improve=15.18952, (0 missing)
## copd < 0.5 to the left, improve=14.68870, (0 missing)
## heart.failure < 0.5 to the left, improve=12.81802, (0 missing)
##
## Node number 202: 3786 observations
## predicted class=B1 expected loss=0.4772847 P(node) =0.01377714
## class counts: 1979 1131 501 157 18
## probabilities: 0.523 0.299 0.132 0.041 0.005
##
## Node number 203: 286 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5839161 P(node) =0.001040746
## class counts: 119 115 36 16 0
## probabilities: 0.416 0.402 0.126 0.056 0.000
## left son=406 (128 obs) right son=407 (158 obs)
## Primary splits:
## age < 73.5 to the left, improve=2.9724540, (0 missing)
## reimbursement2008 < 2005 to the left, improve=1.9802050, (0 missing)
## depression < 0.5 to the left, improve=0.5460014, (0 missing)
## alzheimers < 0.5 to the right, improve=0.4144954, (0 missing)
## ihd < 0.5 to the left, improve=0.3767582, (0 missing)
## Surrogate splits:
## reimbursement2008 < 1945 to the left, agree=0.580, adj=0.063, (0 split)
## arthritis < 0.5 to the right, agree=0.563, adj=0.023, (0 split)
##
## Node number 206: 904 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5376106 P(node) =0.003289629
## class counts: 418 304 119 57 6
## probabilities: 0.462 0.336 0.132 0.063 0.007
## left son=412 (270 obs) right son=413 (634 obs)
## Primary splits:
## reimbursement2008 < 1735 to the left, improve=3.8438620, (0 missing)
## age < 93.5 to the left, improve=3.6681650, (0 missing)
## ihd < 0.5 to the left, improve=3.2669730, (0 missing)
## cancer < 0.5 to the left, improve=3.0869480, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7446912, (0 missing)
## Surrogate splits:
## age < 29 to the left, agree=0.702, adj=0.004, (0 split)
##
## Node number 207: 348 observations
## predicted class=B2 expected loss=0.5890805 P(node) =0.001266362
## class counts: 109 143 63 30 3
## probabilities: 0.313 0.411 0.181 0.086 0.009
##
## Node number 210: 2325 observations
## predicted class=B1 expected loss=0.4541935 P(node) =0.008460606
## class counts: 1269 700 245 99 12
## probabilities: 0.546 0.301 0.105 0.043 0.005
##
## Node number 211: 881 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5471056 P(node) =0.003205933
## class counts: 399 315 118 46 3
## probabilities: 0.453 0.358 0.134 0.052 0.003
## left son=422 (763 obs) right son=423 (118 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=3.122415, (0 missing)
## age < 78.5 to the right, improve=2.656467, (0 missing)
## reimbursement2008 < 2205 to the right, improve=1.600090, (0 missing)
## stroke < 0.5 to the left, improve=1.074836, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.071176, (0 missing)
##
## Node number 212: 1438 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5452017 P(node) =0.00523284
## class counts: 654 515 187 76 6
## probabilities: 0.455 0.358 0.130 0.053 0.004
## left son=424 (495 obs) right son=425 (943 obs)
## Primary splits:
## reimbursement2008 < 2715 to the right, improve=2.835023, (0 missing)
## kidney < 0.5 to the left, improve=1.879898, (0 missing)
## copd < 0.5 to the left, improve=1.857999, (0 missing)
## age < 40.5 to the right, improve=1.802592, (0 missing)
## heart.failure < 0.5 to the left, improve=1.761837, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=0.713, adj=0.166, (0 split)
## age < 28.5 to the left, agree=0.656, adj=0.002, (0 split)
##
## Node number 213: 87 observations
## predicted class=B2 expected loss=0.5517241 P(node) =0.0003165904
## class counts: 27 39 15 6 0
## probabilities: 0.310 0.448 0.172 0.069 0.000
##
## Node number 216: 3992 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.495491 P(node) =0.01452677
## class counts: 2014 1315 497 153 13
## probabilities: 0.505 0.329 0.124 0.038 0.003
## left son=432 (1265 obs) right son=433 (2727 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=7.867939, (0 missing)
## depression < 0.5 to the left, improve=6.016589, (0 missing)
## copd < 0.5 to the left, improve=5.402587, (0 missing)
## kidney < 0.5 to the left, improve=3.916699, (0 missing)
## reimbursement2008 < 2615 to the left, improve=3.836002, (0 missing)
##
## Node number 217: 383 observations, complexity param=0.0001826494
## predicted class=B2 expected loss=0.5744125 P(node) =0.001393726
## class counts: 141 163 58 17 4
## probabilities: 0.368 0.426 0.151 0.044 0.010
## left son=434 (238 obs) right son=435 (145 obs)
## Primary splits:
## reimbursement2008 < 2705 to the left, improve=4.9624930, (0 missing)
## depression < 0.5 to the left, improve=3.2303380, (0 missing)
## age < 67.5 to the right, improve=2.3511250, (0 missing)
## ihd < 0.5 to the left, improve=1.5735720, (0 missing)
## bucket2008 < 1.5 to the left, improve=0.9813303, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.681, adj=0.159, (0 split)
## age < 45.5 to the right, agree=0.624, adj=0.007, (0 split)
##
## Node number 218: 2819 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.5746719 P(node) =0.01025826
## class counts: 1199 980 439 183 18
## probabilities: 0.425 0.348 0.156 0.065 0.006
## left son=436 (635 obs) right son=437 (2184 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=6.072389, (0 missing)
## reimbursement2008 < 2325 to the left, improve=3.797765, (0 missing)
## age < 40.5 to the right, improve=3.110525, (0 missing)
## depression < 0.5 to the left, improve=2.993563, (0 missing)
## stroke < 0.5 to the left, improve=2.412511, (0 missing)
##
## Node number 219: 1219 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.6021329 P(node) =0.004435905
## class counts: 406 485 231 85 12
## probabilities: 0.333 0.398 0.189 0.070 0.010
## left son=438 (613 obs) right son=439 (606 obs)
## Primary splits:
## reimbursement2008 < 2615 to the left, improve=4.2080810, (0 missing)
## age < 98.5 to the right, improve=2.1482090, (0 missing)
## depression < 0.5 to the left, improve=1.6601240, (0 missing)
## stroke < 0.5 to the left, improve=0.8099205, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.7434054, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.579, adj=0.153, (0 split)
## depression < 0.5 to the left, agree=0.523, adj=0.041, (0 split)
## stroke < 0.5 to the left, agree=0.522, adj=0.038, (0 split)
## age < 65.5 to the right, agree=0.519, adj=0.033, (0 split)
## cancer < 0.5 to the left, agree=0.514, adj=0.021, (0 split)
##
## Node number 220: 517 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5299807 P(node) =0.001881348
## class counts: 243 191 57 25 1
## probabilities: 0.470 0.369 0.110 0.048 0.002
## left son=440 (143 obs) right son=441 (374 obs)
## Primary splits:
## reimbursement2008 < 2295 to the left, improve=6.0966680, (0 missing)
## cancer < 0.5 to the left, improve=2.5628030, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.9493160, (0 missing)
## age < 44.5 to the right, improve=1.5968610, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9005685, (0 missing)
## Surrogate splits:
## age < 98.5 to the right, agree=0.729, adj=0.021, (0 split)
##
## Node number 221: 1551 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5963894 P(node) =0.005644043
## class counts: 601 626 223 92 9
## probabilities: 0.387 0.404 0.144 0.059 0.006
## left son=442 (18 obs) right son=443 (1533 obs)
## Primary splits:
## age < 35 to the left, improve=3.0170030, (0 missing)
## kidney < 0.5 to the left, improve=2.3281310, (0 missing)
## cancer < 0.5 to the left, improve=1.5502140, (0 missing)
## stroke < 0.5 to the left, improve=1.1903410, (0 missing)
## copd < 0.5 to the left, improve=0.9727402, (0 missing)
##
## Node number 224: 2292 observations
## predicted class=B1 expected loss=0.3582024 P(node) =0.00834052
## class counts: 1471 549 183 79 10
## probabilities: 0.642 0.240 0.080 0.034 0.004
##
## Node number 225: 843 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5041518 P(node) =0.003067652
## class counts: 418 276 115 34 0
## probabilities: 0.496 0.327 0.136 0.040 0.000
## left son=450 (810 obs) right son=451 (33 obs)
## Primary splits:
## age < 92.5 to the left, improve=5.7055350, (0 missing)
## reimbursement2008 < 11540 to the right, improve=5.6370950, (0 missing)
## bucket2008 < 2.5 to the right, improve=2.9317810, (0 missing)
## stroke < 0.5 to the left, improve=0.7284423, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3506867, (0 missing)
##
## Node number 226: 4266 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.4946085 P(node) =0.01552385
## class counts: 2156 1343 503 238 26
## probabilities: 0.505 0.315 0.118 0.056 0.006
## left son=452 (3304 obs) right son=453 (962 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=10.212680, (0 missing)
## reimbursement2008 < 5905 to the right, improve= 9.673580, (0 missing)
## bucket2008 < 2.5 to the right, improve= 7.844764, (0 missing)
## heart.failure < 0.5 to the left, improve= 6.371374, (0 missing)
## age < 62.5 to the left, improve= 3.683231, (0 missing)
##
## Node number 227: 2224 observations, complexity param=0.0001217663
## predicted class=B1 expected loss=0.6007194 P(node) =0.00809307
## class counts: 888 786 361 169 20
## probabilities: 0.399 0.353 0.162 0.076 0.009
## left son=454 (1518 obs) right son=455 (706 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=6.746609, (0 missing)
## heart.failure < 0.5 to the left, improve=4.569316, (0 missing)
## reimbursement2008 < 10710 to the left, improve=3.711923, (0 missing)
## age < 39.5 to the right, improve=3.285727, (0 missing)
## bucket2008 < 2.5 to the left, improve=2.661027, (0 missing)
## Surrogate splits:
## reimbursement2008 < 14380 to the left, agree=0.714, adj=0.101, (0 split)
## bucket2008 < 3.5 to the left, agree=0.708, adj=0.081, (0 split)
## age < 98.5 to the left, agree=0.684, adj=0.004, (0 split)
##
## Node number 228: 1359 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5548197 P(node) =0.004945361
## class counts: 467 605 203 76 8
## probabilities: 0.344 0.445 0.149 0.056 0.006
## left son=456 (440 obs) right son=457 (919 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.300387, (0 missing)
## alzheimers < 0.5 to the left, improve=2.001264, (0 missing)
## reimbursement2008 < 3265 to the left, improve=1.998939, (0 missing)
## depression < 0.5 to the left, improve=1.755319, (0 missing)
## heart.failure < 0.5 to the left, improve=1.574681, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3095 to the left, agree=0.678, adj=0.007, (0 split)
## age < 29.5 to the left, agree=0.678, adj=0.005, (0 split)
##
## Node number 229: 981 observations
## predicted class=B2 expected loss=0.5259939 P(node) =0.00356983
## class counts: 253 465 188 68 7
## probabilities: 0.258 0.474 0.192 0.069 0.007
##
## Node number 230: 714 observations, complexity param=0.0001881842
## predicted class=B1 expected loss=0.5546218 P(node) =0.002598225
## class counts: 318 239 91 61 5
## probabilities: 0.445 0.335 0.127 0.085 0.007
## left son=460 (412 obs) right son=461 (302 obs)
## Primary splits:
## depression < 0.5 to the left, improve=8.699660, (0 missing)
## age < 92.5 to the right, improve=3.253447, (0 missing)
## reimbursement2008 < 14980 to the left, improve=2.826720, (0 missing)
## bucket2008 < 3.5 to the left, improve=2.191697, (0 missing)
## kidney < 0.5 to the left, improve=2.037790, (0 missing)
## Surrogate splits:
## reimbursement2008 < 32685 to the left, agree=0.591, adj=0.033, (0 split)
## age < 35.5 to the right, agree=0.583, adj=0.013, (0 split)
##
## Node number 231: 444 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.6126126 P(node) =0.001615703
## class counts: 118 172 107 38 9
## probabilities: 0.266 0.387 0.241 0.086 0.020
## left son=462 (282 obs) right son=463 (162 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=3.735228, (0 missing)
## kidney < 0.5 to the left, improve=3.274615, (0 missing)
## reimbursement2008 < 68975 to the right, improve=3.185223, (0 missing)
## ihd < 0.5 to the left, improve=3.085645, (0 missing)
## age < 76.5 to the left, improve=1.652811, (0 missing)
## Surrogate splits:
## age < 95.5 to the left, agree=0.644, adj=0.025, (0 split)
## reimbursement2008 < 8635 to the right, agree=0.637, adj=0.006, (0 split)
##
## Node number 236: 745 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6228188 P(node) =0.002711033
## class counts: 232 281 150 72 10
## probabilities: 0.311 0.377 0.201 0.097 0.013
## left son=472 (159 obs) right son=473 (586 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=3.002920, (0 missing)
## reimbursement2008 < 58135 to the left, improve=2.259882, (0 missing)
## depression < 0.5 to the left, improve=2.111862, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.991400, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.920660, (0 missing)
##
## Node number 237: 309 observations, complexity param=6.167383e-05
## predicted class=B2 expected loss=0.6213592 P(node) =0.001124442
## class counts: 49 117 100 37 6
## probabilities: 0.159 0.379 0.324 0.120 0.019
## left son=474 (237 obs) right son=475 (72 obs)
## Primary splits:
## reimbursement2008 < 10960 to the right, improve=2.966323, (0 missing)
## alzheimers < 0.5 to the right, improve=1.571780, (0 missing)
## age < 90.5 to the left, improve=1.407411, (0 missing)
## copd < 0.5 to the left, improve=1.306020, (0 missing)
## stroke < 0.5 to the left, improve=0.907593, (0 missing)
##
## Node number 238: 243 observations
## predicted class=B2 expected loss=0.526749 P(node) =0.0008842698
## class counts: 33 115 67 24 4
## probabilities: 0.136 0.473 0.276 0.099 0.016
##
## Node number 239: 297 observations, complexity param=6.167383e-05
## predicted class=B3 expected loss=0.6430976 P(node) =0.001080774
## class counts: 52 104 106 33 2
## probabilities: 0.175 0.350 0.357 0.111 0.007
## left son=478 (226 obs) right son=479 (71 obs)
## Primary splits:
## depression < 0.5 to the left, improve=1.480103, (0 missing)
## reimbursement2008 < 6875 to the right, improve=1.383473, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.357727, (0 missing)
## age < 54 to the left, improve=1.263809, (0 missing)
## alzheimers < 0.5 to the right, improve=1.096200, (0 missing)
##
## Node number 240: 2617 observations, complexity param=0.0001439056
## predicted class=B1 expected loss=0.5257929 P(node) =0.009523186
## class counts: 1241 884 351 127 14
## probabilities: 0.474 0.338 0.134 0.049 0.005
## left son=480 (403 obs) right son=481 (2214 obs)
## Primary splits:
## reimbursement2008 < 9400 to the right, improve=12.428110, (0 missing)
## bucket2008 < 2.5 to the right, improve= 8.843694, (0 missing)
## depression < 0.5 to the left, improve= 8.588030, (0 missing)
## osteoporosis < 0.5 to the left, improve= 8.405901, (0 missing)
## alzheimers < 0.5 to the left, improve= 4.036896, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.947, adj=0.658, (0 split)
##
## Node number 241: 9955 observations, complexity param=0.0009879673
## predicted class=B2 expected loss=0.6003014 P(node) =0.03622595
## class counts: 3603 3979 1649 664 60
## probabilities: 0.362 0.400 0.166 0.067 0.006
## left son=482 (5563 obs) right son=483 (4392 obs)
## Primary splits:
## depression < 0.5 to the left, improve=24.69099, (0 missing)
## copd < 0.5 to the left, improve=17.49244, (0 missing)
## heart.failure < 0.5 to the left, improve=17.05734, (0 missing)
## reimbursement2008 < 8955 to the left, improve=14.88623, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.99202, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.574, adj=0.034, (0 split)
## age < 47.5 to the right, agree=0.565, adj=0.013, (0 split)
## copd < 0.5 to the left, agree=0.564, adj=0.013, (0 split)
## reimbursement2008 < 13565 to the left, agree=0.561, adj=0.005, (0 split)
## bucket2008 < 3.5 to the left, agree=0.559, adj=0.001, (0 split)
##
## Node number 244: 4305 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.524971 P(node) =0.01566577
## class counts: 1149 2045 746 328 37
## probabilities: 0.267 0.475 0.173 0.076 0.009
## left son=488 (1063 obs) right son=489 (3242 obs)
## Primary splits:
## reimbursement2008 < 9880 to the right, improve=11.346300, (0 missing)
## bucket2008 < 2.5 to the right, improve= 8.562449, (0 missing)
## ihd < 0.5 to the left, improve= 7.353611, (0 missing)
## copd < 0.5 to the left, improve= 6.701463, (0 missing)
## heart.failure < 0.5 to the left, improve= 3.881008, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.941, adj=0.762, (0 split)
##
## Node number 245: 829 observations
## predicted class=B2 expected loss=0.4885404 P(node) =0.003016707
## class counts: 128 424 190 84 3
## probabilities: 0.154 0.511 0.229 0.101 0.004
##
## Node number 248: 7786 observations, complexity param=0.0001411382
## predicted class=B2 expected loss=0.6050604 P(node) =0.02833302
## class counts: 1982 3075 1667 929 133
## probabilities: 0.255 0.395 0.214 0.119 0.017
## left son=496 (964 obs) right son=497 (6822 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=18.914850, (0 missing)
## depression < 0.5 to the left, improve=16.457650, (0 missing)
## reimbursement2008 < 6325 to the left, improve=12.927220, (0 missing)
## osteoporosis < 0.5 to the left, improve= 9.344273, (0 missing)
## bucket2008 < 2.5 to the left, improve= 9.314433, (0 missing)
##
## Node number 249: 1638 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.005960634
## class counts: 210 702 472 227 27
## probabilities: 0.128 0.429 0.288 0.139 0.016
##
## Node number 252: 3345 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.580568 P(node) =0.01217236
## class counts: 372 1403 837 632 101
## probabilities: 0.111 0.419 0.250 0.189 0.030
## left son=504 (1291 obs) right son=505 (2054 obs)
## Primary splits:
## depression < 0.5 to the left, improve=6.733363, (0 missing)
## copd < 0.5 to the left, improve=6.399894, (0 missing)
## cancer < 0.5 to the left, improve=5.398776, (0 missing)
## heart.failure < 0.5 to the left, improve=3.401421, (0 missing)
## age < 31.5 to the right, improve=3.041832, (0 missing)
## Surrogate splits:
## reimbursement2008 < 15665 to the left, agree=0.614, adj=0.001, (0 split)
##
## Node number 253: 2057 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6212931 P(node) =0.007485362
## class counts: 137 779 473 554 114
## probabilities: 0.067 0.379 0.230 0.269 0.055
## left son=506 (520 obs) right son=507 (1537 obs)
## Primary splits:
## copd < 0.5 to the left, improve=4.741452, (0 missing)
## age < 62.5 to the right, improve=3.709690, (0 missing)
## cancer < 0.5 to the right, improve=3.631891, (0 missing)
## ihd < 0.5 to the left, improve=3.269099, (0 missing)
## heart.failure < 0.5 to the left, improve=3.168350, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the left, agree=0.751, adj=0.015, (0 split)
## age < 29 to the left, agree=0.749, adj=0.008, (0 split)
## ihd < 0.5 to the left, agree=0.749, adj=0.008, (0 split)
##
## Node number 254: 5298 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.689128 P(node) =0.01927927
## class counts: 908 1647 1051 1364 328
## probabilities: 0.171 0.311 0.198 0.257 0.062
## left son=508 (2489 obs) right son=509 (2809 obs)
## Primary splits:
## depression < 0.5 to the left, improve=18.17296, (0 missing)
## reimbursement2008 < 22335 to the left, improve=13.06444, (0 missing)
## copd < 0.5 to the left, improve=11.53148, (0 missing)
## ihd < 0.5 to the left, improve= 8.63716, (0 missing)
## heart.failure < 0.5 to the left, improve= 8.42218, (0 missing)
## Surrogate splits:
## copd < 0.5 to the left, agree=0.579, adj=0.104, (0 split)
## alzheimers < 0.5 to the left, agree=0.573, adj=0.092, (0 split)
## ihd < 0.5 to the left, agree=0.545, adj=0.033, (0 split)
## heart.failure < 0.5 to the left, agree=0.544, adj=0.030, (0 split)
## reimbursement2008 < 16955 to the left, agree=0.535, adj=0.010, (0 split)
##
## Node number 255: 1785 observations
## predicted class=B2 expected loss=0.6369748 P(node) =0.006495562
## class counts: 174 648 492 397 74
## probabilities: 0.097 0.363 0.276 0.222 0.041
##
## Node number 406: 128 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5 P(node) =0.0004657882
## class counts: 64 42 14 8 0
## probabilities: 0.500 0.328 0.109 0.062 0.000
## left son=812 (95 obs) right son=813 (33 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.3313600, (0 missing)
## reimbursement2008 < 2155 to the left, improve=2.1875000, (0 missing)
## age < 70.5 to the right, improve=1.5228130, (0 missing)
## arthritis < 0.5 to the left, improve=1.1806970, (0 missing)
## copd < 0.5 to the left, improve=0.4207762, (0 missing)
##
## Node number 407: 158 observations
## predicted class=B2 expected loss=0.5379747 P(node) =0.0005749573
## class counts: 55 73 22 8 0
## probabilities: 0.348 0.462 0.139 0.051 0.000
##
## Node number 412: 270 observations
## predicted class=B1 expected loss=0.462963 P(node) =0.000982522
## class counts: 145 73 36 15 1
## probabilities: 0.537 0.270 0.133 0.056 0.004
##
## Node number 413: 634 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5694006 P(node) =0.002307107
## class counts: 273 231 83 42 5
## probabilities: 0.431 0.364 0.131 0.066 0.008
## left son=826 (596 obs) right son=827 (38 obs)
## Primary splits:
## age < 91.5 to the left, improve=3.6059530, (0 missing)
## ihd < 0.5 to the left, improve=2.2411130, (0 missing)
## reimbursement2008 < 1765 to the right, improve=2.0115470, (0 missing)
## cancer < 0.5 to the left, improve=1.8824720, (0 missing)
## depression < 0.5 to the right, improve=0.5863526, (0 missing)
##
## Node number 422: 763 observations
## predicted class=B1 expected loss=0.5307995 P(node) =0.002776534
## class counts: 358 260 102 40 3
## probabilities: 0.469 0.341 0.134 0.052 0.004
##
## Node number 423: 118 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5338983 P(node) =0.0004293985
## class counts: 41 55 16 6 0
## probabilities: 0.347 0.466 0.136 0.051 0.000
## left son=846 (22 obs) right son=847 (96 obs)
## Primary splits:
## reimbursement2008 < 2865 to the right, improve=2.6611130, (0 missing)
## copd < 0.5 to the left, improve=1.5528850, (0 missing)
## heart.failure < 0.5 to the left, improve=1.3108310, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.2553930, (0 missing)
## age < 89.5 to the left, improve=0.9696791, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=0.873, adj=0.318, (0 split)
##
## Node number 424: 495 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5656566 P(node) =0.00180129
## class counts: 215 202 50 26 2
## probabilities: 0.434 0.408 0.101 0.053 0.004
## left son=848 (385 obs) right son=849 (110 obs)
## Primary splits:
## reimbursement2008 < 2795 to the right, improve=2.2427130, (0 missing)
## age < 73.5 to the left, improve=1.5903260, (0 missing)
## ihd < 0.5 to the left, improve=1.1717170, (0 missing)
## depression < 0.5 to the left, improve=0.5724615, (0 missing)
## kidney < 0.5 to the left, improve=0.5572971, (0 missing)
##
## Node number 425: 943 observations
## predicted class=B1 expected loss=0.5344645 P(node) =0.003431549
## class counts: 439 313 137 50 4
## probabilities: 0.466 0.332 0.145 0.053 0.004
##
## Node number 432: 1265 observations
## predicted class=B1 expected loss=0.4442688 P(node) =0.004603298
## class counts: 703 367 147 44 4
## probabilities: 0.556 0.290 0.116 0.035 0.003
##
## Node number 433: 2727 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5192519 P(node) =0.009923472
## class counts: 1311 948 350 109 9
## probabilities: 0.481 0.348 0.128 0.040 0.003
## left son=866 (1499 obs) right son=867 (1228 obs)
## Primary splits:
## reimbursement2008 < 2615 to the left, improve=4.028460, (0 missing)
## age < 54.5 to the left, improve=3.426946, (0 missing)
## copd < 0.5 to the left, improve=2.215284, (0 missing)
## stroke < 0.5 to the left, improve=2.121876, (0 missing)
## depression < 0.5 to the left, improve=1.918448, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.612, adj=0.139, (0 split)
## age < 97.5 to the left, agree=0.552, adj=0.005, (0 split)
##
## Node number 434: 238 observations, complexity param=0.0001826494
## predicted class=B1 expected loss=0.5714286 P(node) =0.000866075
## class counts: 102 86 38 9 3
## probabilities: 0.429 0.361 0.160 0.038 0.013
## left son=868 (167 obs) right son=869 (71 obs)
## Primary splits:
## depression < 0.5 to the left, improve=4.834875, (0 missing)
## age < 59.5 to the right, improve=2.461134, (0 missing)
## ihd < 0.5 to the left, improve=1.909944, (0 missing)
## reimbursement2008 < 2285 to the left, improve=1.842456, (0 missing)
## alzheimers < 0.5 to the left, improve=1.113912, (0 missing)
## Surrogate splits:
## age < 97.5 to the left, agree=0.718, adj=0.056, (0 split)
##
## Node number 435: 145 observations
## predicted class=B2 expected loss=0.4689655 P(node) =0.0005276507
## class counts: 39 77 20 8 1
## probabilities: 0.269 0.531 0.138 0.055 0.007
##
## Node number 436: 635 observations
## predicted class=B1 expected loss=0.5023622 P(node) =0.002310746
## class counts: 316 196 93 26 4
## probabilities: 0.498 0.309 0.146 0.041 0.006
##
## Node number 437: 2184 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.595696 P(node) =0.007947511
## class counts: 883 784 346 157 14
## probabilities: 0.404 0.359 0.158 0.072 0.006
## left son=874 (393 obs) right son=875 (1791 obs)
## Primary splits:
## reimbursement2008 < 2315 to the left, improve=4.386891, (0 missing)
## depression < 0.5 to the left, improve=4.376862, (0 missing)
## age < 39.5 to the right, improve=3.004733, (0 missing)
## alzheimers < 0.5 to the left, improve=2.391734, (0 missing)
## stroke < 0.5 to the left, improve=2.171601, (0 missing)
##
## Node number 438: 613 observations, complexity param=8.302246e-05
## predicted class=B1 expected loss=0.6182708 P(node) =0.002230689
## class counts: 234 226 111 36 6
## probabilities: 0.382 0.369 0.181 0.059 0.010
## left son=876 (180 obs) right son=877 (433 obs)
## Primary splits:
## osteoporosis < 0.5 to the right, improve=1.4494640, (0 missing)
## age < 98.5 to the right, improve=1.3979840, (0 missing)
## stroke < 0.5 to the left, improve=0.9190213, (0 missing)
## reimbursement2008 < 2275 to the left, improve=0.8284921, (0 missing)
## depression < 0.5 to the left, improve=0.7804891, (0 missing)
## Surrogate splits:
## reimbursement2008 < 2605 to the right, agree=0.713, adj=0.022, (0 split)
## age < 99.5 to the right, agree=0.711, adj=0.017, (0 split)
##
## Node number 439: 606 observations
## predicted class=B2 expected loss=0.5726073 P(node) =0.002205216
## class counts: 172 259 120 49 6
## probabilities: 0.284 0.427 0.198 0.081 0.010
##
## Node number 440: 143 observations
## predicted class=B1 expected loss=0.3986014 P(node) =0.0005203728
## class counts: 86 37 11 9 0
## probabilities: 0.601 0.259 0.077 0.063 0.000
##
## Node number 441: 374 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5802139 P(node) =0.001360975
## class counts: 157 154 46 16 1
## probabilities: 0.420 0.412 0.123 0.043 0.003
## left son=882 (25 obs) right son=883 (349 obs)
## Primary splits:
## reimbursement2008 < 2315 to the left, improve=4.569334, (0 missing)
## cancer < 0.5 to the left, improve=2.355946, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.361181, (0 missing)
## age < 90.5 to the right, improve=1.103565, (0 missing)
## heart.failure < 0.5 to the left, improve=1.082873, (0 missing)
##
## Node number 442: 18 observations
## predicted class=B1 expected loss=0.2777778 P(node) =6.550147e-05
## class counts: 13 4 0 1 0
## probabilities: 0.722 0.222 0.000 0.056 0.000
##
## Node number 443: 1533 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5942596 P(node) =0.005578542
## class counts: 588 622 223 91 9
## probabilities: 0.384 0.406 0.145 0.059 0.006
## left son=886 (1101 obs) right son=887 (432 obs)
## Primary splits:
## kidney < 0.5 to the left, improve=2.2490350, (0 missing)
## cancer < 0.5 to the left, improve=1.4724050, (0 missing)
## stroke < 0.5 to the left, improve=1.3260620, (0 missing)
## reimbursement2008 < 2435 to the left, improve=1.1404580, (0 missing)
## copd < 0.5 to the left, improve=0.9660973, (0 missing)
##
## Node number 450: 810 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.491358 P(node) =0.002947566
## class counts: 412 257 108 33 0
## probabilities: 0.509 0.317 0.133 0.041 0.000
## left son=900 (117 obs) right son=901 (693 obs)
## Primary splits:
## reimbursement2008 < 11525 to the right, improve=5.1504220, (0 missing)
## bucket2008 < 2.5 to the right, improve=2.8801500, (0 missing)
## age < 34.5 to the left, improve=1.2970260, (0 missing)
## stroke < 0.5 to the left, improve=0.8677994, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5279869, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.928, adj=0.504, (0 split)
##
## Node number 451: 33 observations
## predicted class=B2 expected loss=0.4242424 P(node) =0.000120086
## class counts: 6 19 7 1 0
## probabilities: 0.182 0.576 0.212 0.030 0.000
##
## Node number 452: 3304 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.4757869 P(node) =0.01202316
## class counts: 1732 979 389 183 21
## probabilities: 0.524 0.296 0.118 0.055 0.006
## left son=904 (1626 obs) right son=905 (1678 obs)
## Primary splits:
## reimbursement2008 < 5905 to the right, improve=9.971499, (0 missing)
## bucket2008 < 2.5 to the right, improve=7.851328, (0 missing)
## age < 62.5 to the left, improve=4.441278, (0 missing)
## heart.failure < 0.5 to the left, improve=3.580749, (0 missing)
## kidney < 0.5 to the left, improve=1.765354, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.861, adj=0.717, (0 split)
## kidney < 0.5 to the right, agree=0.637, adj=0.262, (0 split)
## heart.failure < 0.5 to the right, agree=0.608, adj=0.204, (0 split)
## copd < 0.5 to the right, agree=0.590, adj=0.166, (0 split)
## alzheimers < 0.5 to the right, agree=0.558, adj=0.102, (0 split)
##
## Node number 453: 962 observations, complexity param=0.0001051618
## predicted class=B1 expected loss=0.5592516 P(node) =0.00350069
## class counts: 424 364 114 55 5
## probabilities: 0.441 0.378 0.119 0.057 0.005
## left son=906 (857 obs) right son=907 (105 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=3.576264, (0 missing)
## heart.failure < 0.5 to the left, improve=3.114700, (0 missing)
## reimbursement2008 < 59635 to the left, improve=2.145451, (0 missing)
## age < 97.5 to the right, improve=1.742305, (0 missing)
## copd < 0.5 to the left, improve=1.012750, (0 missing)
##
## Node number 454: 1518 observations
## predicted class=B1 expected loss=0.5685112 P(node) =0.005523957
## class counts: 655 520 243 92 8
## probabilities: 0.431 0.343 0.160 0.061 0.005
##
## Node number 455: 706 observations, complexity param=0.0001162314
## predicted class=B2 expected loss=0.6232295 P(node) =0.002569113
## class counts: 233 266 118 77 12
## probabilities: 0.330 0.377 0.167 0.109 0.017
## left son=910 (696 obs) right son=911 (10 obs)
## Primary splits:
## reimbursement2008 < 3155 to the right, improve=3.301177, (0 missing)
## heart.failure < 0.5 to the left, improve=3.232296, (0 missing)
## copd < 0.5 to the left, improve=2.330270, (0 missing)
## alzheimers < 0.5 to the left, improve=1.835216, (0 missing)
## age < 92.5 to the right, improve=1.805094, (0 missing)
##
## Node number 456: 440 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5636364 P(node) =0.001601147
## class counts: 177 192 49 20 2
## probabilities: 0.402 0.436 0.111 0.045 0.005
## left son=912 (58 obs) right son=913 (382 obs)
## Primary splits:
## reimbursement2008 < 3155 to the left, improve=3.3827400, (0 missing)
## age < 80.5 to the right, improve=2.1481460, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6300393, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.5745512, (0 missing)
## depression < 0.5 to the right, improve=0.5547491, (0 missing)
##
## Node number 457: 919 observations
## predicted class=B2 expected loss=0.5505985 P(node) =0.003344214
## class counts: 290 413 154 56 6
## probabilities: 0.316 0.449 0.168 0.061 0.007
##
## Node number 460: 412 observations
## predicted class=B1 expected loss=0.4757282 P(node) =0.001499256
## class counts: 216 120 42 30 4
## probabilities: 0.524 0.291 0.102 0.073 0.010
##
## Node number 461: 302 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6059603 P(node) =0.001098969
## class counts: 102 119 49 31 1
## probabilities: 0.338 0.394 0.162 0.103 0.003
## left son=922 (9 obs) right son=923 (293 obs)
## Primary splits:
## age < 92.5 to the right, improve=2.5766490, (0 missing)
## stroke < 0.5 to the right, improve=1.8961920, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.0608030, (0 missing)
## reimbursement2008 < 32980 to the right, improve=1.0319510, (0 missing)
## kidney < 0.5 to the left, improve=0.7951977, (0 missing)
##
## Node number 462: 282 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.6631206 P(node) =0.00102619
## class counts: 88 95 71 23 5
## probabilities: 0.312 0.337 0.252 0.082 0.018
## left son=924 (220 obs) right son=925 (62 obs)
## Primary splits:
## reimbursement2008 < 27390 to the left, improve=2.933452, (0 missing)
## age < 79.5 to the left, improve=2.171675, (0 missing)
## bucket2008 < 4.5 to the right, improve=1.933271, (0 missing)
## kidney < 0.5 to the left, improve=1.142914, (0 missing)
## ihd < 0.5 to the left, improve=1.125301, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.816, adj=0.161, (0 split)
##
## Node number 463: 162 observations
## predicted class=B2 expected loss=0.5246914 P(node) =0.0005895132
## class counts: 30 77 36 15 4
## probabilities: 0.185 0.475 0.222 0.093 0.025
##
## Node number 472: 159 observations, complexity param=6.167383e-05
## predicted class=B1 expected loss=0.591195 P(node) =0.0005785963
## class counts: 65 51 33 8 2
## probabilities: 0.409 0.321 0.208 0.050 0.013
## left son=944 (76 obs) right son=945 (83 obs)
## Primary splits:
## reimbursement2008 < 11995 to the right, improve=3.4294220, (0 missing)
## age < 65 to the left, improve=1.4674530, (0 missing)
## copd < 0.5 to the left, improve=1.0021090, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7722947, (0 missing)
## bucket2008 < 3.5 to the left, improve=0.4091195, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.673, adj=0.316, (0 split)
## alzheimers < 0.5 to the right, agree=0.591, adj=0.145, (0 split)
## copd < 0.5 to the right, agree=0.572, adj=0.105, (0 split)
## heart.failure < 0.5 to the right, agree=0.560, adj=0.079, (0 split)
## age < 85.5 to the right, agree=0.541, adj=0.039, (0 split)
##
## Node number 473: 586 observations
## predicted class=B2 expected loss=0.6075085 P(node) =0.002132437
## class counts: 167 230 117 64 8
## probabilities: 0.285 0.392 0.200 0.109 0.014
##
## Node number 474: 237 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5738397 P(node) =0.000862436
## class counts: 35 101 72 26 3
## probabilities: 0.148 0.426 0.304 0.110 0.013
## left son=948 (126 obs) right son=949 (111 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.2728500, (0 missing)
## reimbursement2008 < 18275 to the left, improve=1.9530690, (0 missing)
## bucket2008 < 3.5 to the left, improve=1.1622440, (0 missing)
## age < 75.5 to the right, improve=1.0471110, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8993424, (0 missing)
## Surrogate splits:
## age < 86.5 to the left, agree=0.599, adj=0.144, (0 split)
## heart.failure < 0.5 to the left, agree=0.586, adj=0.117, (0 split)
## kidney < 0.5 to the left, agree=0.582, adj=0.108, (0 split)
## depression < 0.5 to the left, agree=0.570, adj=0.081, (0 split)
## stroke < 0.5 to the left, agree=0.565, adj=0.072, (0 split)
##
## Node number 475: 72 observations
## predicted class=B3 expected loss=0.6111111 P(node) =0.0002620059
## class counts: 14 16 28 11 3
## probabilities: 0.194 0.222 0.389 0.153 0.042
##
## Node number 478: 226 observations
## predicted class=B2 expected loss=0.6238938 P(node) =0.0008224073
## class counts: 40 85 74 26 1
## probabilities: 0.177 0.376 0.327 0.115 0.004
##
## Node number 479: 71 observations
## predicted class=B3 expected loss=0.5492958 P(node) =0.0002583669
## class counts: 12 19 32 7 1
## probabilities: 0.169 0.268 0.451 0.099 0.014
##
## Node number 480: 403 observations
## predicted class=B1 expected loss=0.4243176 P(node) =0.001466505
## class counts: 232 86 61 20 4
## probabilities: 0.576 0.213 0.151 0.050 0.010
##
## Node number 481: 2214 observations, complexity param=0.0001439056
## predicted class=B1 expected loss=0.5442638 P(node) =0.008056681
## class counts: 1009 798 290 107 10
## probabilities: 0.456 0.360 0.131 0.048 0.005
## left son=962 (1636 obs) right son=963 (578 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=6.829533, (0 missing)
## depression < 0.5 to the left, improve=5.940363, (0 missing)
## copd < 0.5 to the left, improve=3.544519, (0 missing)
## alzheimers < 0.5 to the left, improve=2.874488, (0 missing)
## age < 57.5 to the left, improve=2.182371, (0 missing)
## Surrogate splits:
## reimbursement2008 < 9185 to the left, agree=0.742, adj=0.01, (0 split)
##
## Node number 482: 5563 observations, complexity param=0.000785946
## predicted class=B1 expected loss=0.6002157 P(node) =0.02024359
## class counts: 2224 2125 853 328 33
## probabilities: 0.400 0.382 0.153 0.059 0.006
## left son=964 (1363 obs) right son=965 (4200 obs)
## Primary splits:
## reimbursement2008 < 8955 to the right, improve=10.881900, (0 missing)
## heart.failure < 0.5 to the left, improve= 9.391652, (0 missing)
## copd < 0.5 to the left, improve= 8.088751, (0 missing)
## bucket2008 < 2.5 to the right, improve= 7.479912, (0 missing)
## age < 31.5 to the left, improve= 2.090877, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.959, adj=0.834, (0 split)
## age < 28.5 to the left, agree=0.755, adj=0.001, (0 split)
##
## Node number 483: 4392 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5778689 P(node) =0.01598236
## class counts: 1379 1854 796 336 27
## probabilities: 0.314 0.422 0.181 0.077 0.006
## left son=966 (2928 obs) right son=967 (1464 obs)
## Primary splits:
## reimbursement2008 < 8325 to the left, improve=7.509791, (0 missing)
## copd < 0.5 to the left, improve=6.714633, (0 missing)
## heart.failure < 0.5 to the left, improve=5.287260, (0 missing)
## bucket2008 < 2.5 to the left, improve=5.126640, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.940785, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.979, adj=0.936, (0 split)
##
## Node number 488: 1063 observations, complexity param=6.918538e-05
## predicted class=B2 expected loss=0.5983067 P(node) =0.003868226
## class counts: 336 427 192 95 13
## probabilities: 0.316 0.402 0.181 0.089 0.012
## left son=976 (102 obs) right son=977 (961 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=6.866573, (0 missing)
## heart.failure < 0.5 to the left, improve=4.371067, (0 missing)
## copd < 0.5 to the left, improve=3.836869, (0 missing)
## reimbursement2008 < 18130 to the left, improve=3.632764, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.933358, (0 missing)
##
## Node number 489: 3242 observations
## predicted class=B2 expected loss=0.5009254 P(node) =0.01179754
## class counts: 813 1618 554 233 24
## probabilities: 0.251 0.499 0.171 0.072 0.007
##
## Node number 496: 964 observations, complexity param=0.0001411382
## predicted class=B1 expected loss=0.6307054 P(node) =0.003507968
## class counts: 356 348 167 82 11
## probabilities: 0.369 0.361 0.173 0.085 0.011
## left son=992 (572 obs) right son=993 (392 obs)
## Primary splits:
## depression < 0.5 to the left, improve=5.116701, (0 missing)
## reimbursement2008 < 8255 to the left, improve=4.417859, (0 missing)
## bucket2008 < 2.5 to the left, improve=3.940989, (0 missing)
## heart.failure < 0.5 to the left, improve=3.162605, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.038248, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.601, adj=0.018, (0 split)
## reimbursement2008 < 14855 to the left, agree=0.598, adj=0.010, (0 split)
## copd < 0.5 to the left, agree=0.594, adj=0.003, (0 split)
##
## Node number 497: 6822 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6002639 P(node) =0.02482506
## class counts: 1626 2727 1500 847 122
## probabilities: 0.238 0.400 0.220 0.124 0.018
## left son=994 (3172 obs) right son=995 (3650 obs)
## Primary splits:
## reimbursement2008 < 6325 to the left, improve=11.481700, (0 missing)
## depression < 0.5 to the left, improve=11.166950, (0 missing)
## bucket2008 < 2.5 to the left, improve= 7.602059, (0 missing)
## osteoporosis < 0.5 to the left, improve= 6.404955, (0 missing)
## copd < 0.5 to the left, improve= 5.945024, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.869, adj=0.717, (0 split)
## copd < 0.5 to the left, agree=0.576, adj=0.088, (0 split)
## heart.failure < 0.5 to the left, agree=0.570, adj=0.076, (0 split)
## alzheimers < 0.5 to the left, agree=0.545, adj=0.020, (0 split)
## age < 31.5 to the left, agree=0.536, adj=0.003, (0 split)
##
## Node number 504: 1291 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5530596 P(node) =0.004697911
## class counts: 183 577 280 219 32
## probabilities: 0.142 0.447 0.217 0.170 0.025
## left son=1008 (973 obs) right son=1009 (318 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=5.737334, (0 missing)
## reimbursement2008 < 32055 to the left, improve=1.892346, (0 missing)
## age < 84.5 to the right, improve=1.763761, (0 missing)
## ihd < 0.5 to the left, improve=1.710826, (0 missing)
## alzheimers < 0.5 to the right, improve=1.100420, (0 missing)
## Surrogate splits:
## age < 28.5 to the right, agree=0.755, adj=0.006, (0 split)
##
## Node number 505: 2054 observations
## predicted class=B2 expected loss=0.5978578 P(node) =0.007474445
## class counts: 189 826 557 413 69
## probabilities: 0.092 0.402 0.271 0.201 0.034
##
## Node number 506: 520 observations
## predicted class=B2 expected loss=0.5769231 P(node) =0.001892265
## class counts: 50 220 120 108 22
## probabilities: 0.096 0.423 0.231 0.208 0.042
##
## Node number 507: 1537 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6363045 P(node) =0.005593098
## class counts: 87 559 353 446 92
## probabilities: 0.057 0.364 0.230 0.290 0.060
## left son=1014 (1286 obs) right son=1015 (251 obs)
## Primary splits:
## age < 62.5 to the right, improve=4.292573, (0 missing)
## reimbursement2008 < 43950 to the left, improve=3.358938, (0 missing)
## cancer < 0.5 to the right, improve=2.803709, (0 missing)
## heart.failure < 0.5 to the left, improve=1.956332, (0 missing)
## stroke < 0.5 to the left, improve=1.605851, (0 missing)
##
## Node number 508: 2489 observations, complexity param=0.0002036818
## predicted class=B2 expected loss=0.7219767 P(node) =0.009057397
## class counts: 541 692 436 670 150
## probabilities: 0.217 0.278 0.175 0.269 0.060
## left son=1016 (1317 obs) right son=1017 (1172 obs)
## Primary splits:
## copd < 0.5 to the right, improve=9.293163, (0 missing)
## reimbursement2008 < 23175 to the left, improve=7.265866, (0 missing)
## ihd < 0.5 to the left, improve=7.177016, (0 missing)
## heart.failure < 0.5 to the left, improve=4.187307, (0 missing)
## bucket2008 < 4.5 to the left, improve=3.684093, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the right, agree=0.575, adj=0.097, (0 split)
## alzheimers < 0.5 to the right, agree=0.556, adj=0.057, (0 split)
## ihd < 0.5 to the right, agree=0.555, adj=0.055, (0 split)
## reimbursement2008 < 27380 to the right, agree=0.544, adj=0.032, (0 split)
## age < 52.5 to the right, agree=0.532, adj=0.005, (0 split)
##
## Node number 509: 2809 observations
## predicted class=B2 expected loss=0.6600214 P(node) =0.01022187
## class counts: 367 955 615 694 178
## probabilities: 0.131 0.340 0.219 0.247 0.063
##
## Node number 812: 95 observations
## predicted class=B1 expected loss=0.4315789 P(node) =0.0003457022
## class counts: 54 25 11 5 0
## probabilities: 0.568 0.263 0.116 0.053 0.000
##
## Node number 813: 33 observations
## predicted class=B2 expected loss=0.4848485 P(node) =0.000120086
## class counts: 10 17 3 3 0
## probabilities: 0.303 0.515 0.091 0.091 0.000
##
## Node number 826: 596 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5536913 P(node) =0.002168826
## class counts: 266 214 77 34 5
## probabilities: 0.446 0.359 0.129 0.057 0.008
## left son=1652 (555 obs) right son=1653 (41 obs)
## Primary splits:
## reimbursement2008 < 1765 to the right, improve=2.4482060, (0 missing)
## age < 81.5 to the right, improve=2.3548750, (0 missing)
## ihd < 0.5 to the left, improve=2.3213280, (0 missing)
## cancer < 0.5 to the left, improve=2.1512770, (0 missing)
## depression < 0.5 to the left, improve=0.5755867, (0 missing)
##
## Node number 827: 38 observations
## predicted class=B2 expected loss=0.5526316 P(node) =0.0001382809
## class counts: 7 17 6 8 0
## probabilities: 0.184 0.447 0.158 0.211 0.000
##
## Node number 846: 22 observations
## predicted class=B1 expected loss=0.4545455 P(node) =8.005735e-05
## class counts: 12 5 4 1 0
## probabilities: 0.545 0.227 0.182 0.045 0.000
##
## Node number 847: 96 observations
## predicted class=B2 expected loss=0.4791667 P(node) =0.0003493412
## class counts: 29 50 12 5 0
## probabilities: 0.302 0.521 0.125 0.052 0.000
##
## Node number 848: 385 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5454545 P(node) =0.001401004
## class counts: 175 146 39 23 2
## probabilities: 0.455 0.379 0.101 0.060 0.005
## left son=1696 (263 obs) right son=1697 (122 obs)
## Primary splits:
## age < 80.5 to the left, improve=2.5496070, (0 missing)
## ihd < 0.5 to the left, improve=1.7465940, (0 missing)
## reimbursement2008 < 3025 to the right, improve=0.9395484, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7038989, (0 missing)
## depression < 0.5 to the left, improve=0.3133237, (0 missing)
##
## Node number 849: 110 observations
## predicted class=B2 expected loss=0.4909091 P(node) =0.0004002868
## class counts: 40 56 11 3 0
## probabilities: 0.364 0.509 0.100 0.027 0.000
##
## Node number 866: 1499 observations
## predicted class=B1 expected loss=0.490994 P(node) =0.005454817
## class counts: 763 492 189 52 3
## probabilities: 0.509 0.328 0.126 0.035 0.002
##
## Node number 867: 1228 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5537459 P(node) =0.004468656
## class counts: 548 456 161 57 6
## probabilities: 0.446 0.371 0.131 0.046 0.005
## left son=1734 (171 obs) right son=1735 (1057 obs)
## Primary splits:
## reimbursement2008 < 2995 to the right, improve=3.399761, (0 missing)
## bucket2008 < 1.5 to the right, improve=3.399761, (0 missing)
## age < 83.5 to the left, improve=2.697325, (0 missing)
## kidney < 0.5 to the left, improve=2.465389, (0 missing)
## depression < 0.5 to the left, improve=1.722272, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the right, agree=1, adj=1, (0 split)
##
## Node number 868: 167 observations
## predicted class=B1 expected loss=0.502994 P(node) =0.0006077081
## class counts: 83 50 25 7 2
## probabilities: 0.497 0.299 0.150 0.042 0.012
##
## Node number 869: 71 observations
## predicted class=B2 expected loss=0.4929577 P(node) =0.0002583669
## class counts: 19 36 13 2 1
## probabilities: 0.268 0.507 0.183 0.028 0.014
##
## Node number 874: 393 observations
## predicted class=B1 expected loss=0.5139949 P(node) =0.001430115
## class counts: 191 134 46 20 2
## probabilities: 0.486 0.341 0.117 0.051 0.005
##
## Node number 875: 1791 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.6136237 P(node) =0.006517396
## class counts: 692 650 300 137 12
## probabilities: 0.386 0.363 0.168 0.076 0.007
## left son=1750 (1752 obs) right son=1751 (39 obs)
## Primary splits:
## age < 39.5 to the right, improve=3.631907, (0 missing)
## depression < 0.5 to the left, improve=3.598994, (0 missing)
## reimbursement2008 < 2475 to the left, improve=1.828499, (0 missing)
## alzheimers < 0.5 to the left, improve=1.790378, (0 missing)
## stroke < 0.5 to the left, improve=1.609073, (0 missing)
##
## Node number 876: 180 observations, complexity param=8.302246e-05
## predicted class=B1 expected loss=0.5666667 P(node) =0.0006550147
## class counts: 78 68 23 11 0
## probabilities: 0.433 0.378 0.128 0.061 0.000
## left son=1752 (112 obs) right son=1753 (68 obs)
## Primary splits:
## reimbursement2008 < 2455 to the left, improve=3.4787820, (0 missing)
## age < 88.5 to the left, improve=1.3486290, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8074074, (0 missing)
## copd < 0.5 to the left, improve=0.6952328, (0 missing)
## ihd < 0.5 to the right, improve=0.3305556, (0 missing)
##
## Node number 877: 433 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.6351039 P(node) =0.001575674
## class counts: 156 158 88 25 6
## probabilities: 0.360 0.365 0.203 0.058 0.014
## left son=1754 (403 obs) right son=1755 (30 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=1.8988510, (0 missing)
## age < 45.5 to the left, improve=1.3010460, (0 missing)
## reimbursement2008 < 2255 to the left, improve=1.2799960, (0 missing)
## depression < 0.5 to the left, improve=1.2338070, (0 missing)
## cancer < 0.5 to the right, improve=0.6525541, (0 missing)
##
## Node number 882: 25 observations
## predicted class=B2 expected loss=0.24 P(node) =9.097426e-05
## class counts: 6 19 0 0 0
## probabilities: 0.240 0.760 0.000 0.000 0.000
##
## Node number 883: 349 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5673352 P(node) =0.001270001
## class counts: 151 135 46 16 1
## probabilities: 0.433 0.387 0.132 0.046 0.003
## left son=1766 (336 obs) right son=1767 (13 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.2574550, (0 missing)
## age < 69.5 to the left, improve=1.5047970, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.0520090, (0 missing)
## heart.failure < 0.5 to the left, improve=0.8771249, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.8193194, (0 missing)
##
## Node number 886: 1101 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.595822 P(node) =0.004006506
## class counts: 445 444 150 57 5
## probabilities: 0.404 0.403 0.136 0.052 0.005
## left son=1772 (1057 obs) right son=1773 (44 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=2.0669290, (0 missing)
## age < 46.5 to the left, improve=1.0570040, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.9632398, (0 missing)
## cancer < 0.5 to the left, improve=0.9197684, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7998904, (0 missing)
##
## Node number 887: 432 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.587963 P(node) =0.001572035
## class counts: 143 178 73 34 4
## probabilities: 0.331 0.412 0.169 0.079 0.009
## left son=1774 (403 obs) right son=1775 (29 obs)
## Primary splits:
## reimbursement2008 < 2215 to the right, improve=1.979831, (0 missing)
## depression < 0.5 to the left, improve=1.399095, (0 missing)
## age < 65.5 to the right, improve=1.336452, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.190308, (0 missing)
## heart.failure < 0.5 to the left, improve=1.062301, (0 missing)
##
## Node number 900: 117 observations
## predicted class=B1 expected loss=0.3418803 P(node) =0.0004257595
## class counts: 77 25 8 7 0
## probabilities: 0.658 0.214 0.068 0.060 0.000
##
## Node number 901: 693 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5165945 P(node) =0.002521807
## class counts: 335 232 100 26 0
## probabilities: 0.483 0.335 0.144 0.038 0.000
## left son=1802 (684 obs) right son=1803 (9 obs)
## Primary splits:
## reimbursement2008 < 11105 to the left, improve=2.2165640, (0 missing)
## alzheimers < 0.5 to the left, improve=1.2536740, (0 missing)
## stroke < 0.5 to the left, improve=0.8354877, (0 missing)
## age < 34 to the left, improve=0.8168384, (0 missing)
## copd < 0.5 to the left, improve=0.3325841, (0 missing)
##
## Node number 904: 1626 observations
## predicted class=B1 expected loss=0.4391144 P(node) =0.005916966
## class counts: 912 414 190 99 11
## probabilities: 0.561 0.255 0.117 0.061 0.007
##
## Node number 905: 1678 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.511323 P(node) =0.006106192
## class counts: 820 565 199 84 10
## probabilities: 0.489 0.337 0.119 0.050 0.006
## left son=1810 (1608 obs) right son=1811 (70 obs)
## Primary splits:
## reimbursement2008 < 5695 to the left, improve=3.4228850, (0 missing)
## age < 54.5 to the left, improve=3.4011680, (0 missing)
## kidney < 0.5 to the left, improve=2.7930360, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6256038, (0 missing)
## stroke < 0.5 to the left, improve=0.4872841, (0 missing)
##
## Node number 906: 857 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.5425904 P(node) =0.003118598
## class counts: 392 313 100 48 4
## probabilities: 0.457 0.365 0.117 0.056 0.005
## left son=1812 (405 obs) right son=1813 (452 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=1.8446560, (0 missing)
## reimbursement2008 < 8165 to the right, improve=1.5511950, (0 missing)
## age < 58.5 to the right, improve=1.3750490, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.0178390, (0 missing)
## kidney < 0.5 to the left, improve=0.9078092, (0 missing)
## Surrogate splits:
## reimbursement2008 < 7010 to the left, agree=0.629, adj=0.215, (0 split)
## bucket2008 < 2.5 to the left, agree=0.611, adj=0.178, (0 split)
## kidney < 0.5 to the left, agree=0.585, adj=0.121, (0 split)
## copd < 0.5 to the left, agree=0.583, adj=0.119, (0 split)
## age < 77.5 to the left, agree=0.555, adj=0.059, (0 split)
##
## Node number 907: 105 observations
## predicted class=B2 expected loss=0.5142857 P(node) =0.0003820919
## class counts: 32 51 14 7 1
## probabilities: 0.305 0.486 0.133 0.067 0.010
##
## Node number 910: 696 observations, complexity param=0.0001162314
## predicted class=B2 expected loss=0.6192529 P(node) =0.002532723
## class counts: 232 265 112 75 12
## probabilities: 0.333 0.381 0.161 0.108 0.017
## left son=1820 (177 obs) right son=1821 (519 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.566495, (0 missing)
## copd < 0.5 to the left, improve=2.688595, (0 missing)
## age < 70.5 to the right, improve=1.873250, (0 missing)
## alzheimers < 0.5 to the left, improve=1.737201, (0 missing)
## reimbursement2008 < 7405 to the left, improve=1.699902, (0 missing)
##
## Node number 911: 10 observations
## predicted class=B3 expected loss=0.4 P(node) =3.63897e-05
## class counts: 1 1 6 2 0
## probabilities: 0.100 0.100 0.600 0.200 0.000
##
## Node number 912: 58 observations
## predicted class=B2 expected loss=0.3793103 P(node) =0.0002110603
## class counts: 20 36 0 1 1
## probabilities: 0.345 0.621 0.000 0.017 0.017
##
## Node number 913: 382 observations, complexity param=5.313437e-05
## predicted class=B1 expected loss=0.5890052 P(node) =0.001390087
## class counts: 157 156 49 19 1
## probabilities: 0.411 0.408 0.128 0.050 0.003
## left son=1826 (25 obs) right son=1827 (357 obs)
## Primary splits:
## reimbursement2008 < 3245 to the left, improve=2.6514540, (0 missing)
## age < 80.5 to the right, improve=2.1655330, (0 missing)
## depression < 0.5 to the left, improve=1.2095670, (0 missing)
## kidney < 0.5 to the left, improve=1.1024610, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7876318, (0 missing)
##
## Node number 922: 9 observations
## predicted class=B1 expected loss=0.3333333 P(node) =3.275073e-05
## class counts: 6 0 1 1 1
## probabilities: 0.667 0.000 0.111 0.111 0.111
##
## Node number 923: 293 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.5938567 P(node) =0.001066218
## class counts: 96 119 48 30 0
## probabilities: 0.328 0.406 0.164 0.102 0.000
## left son=1846 (39 obs) right son=1847 (254 obs)
## Primary splits:
## stroke < 0.5 to the right, improve=2.1766940, (0 missing)
## age < 65.5 to the right, improve=1.5636490, (0 missing)
## reimbursement2008 < 11660 to the right, improve=1.0372370, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9901623, (0 missing)
## kidney < 0.5 to the left, improve=0.7188410, (0 missing)
## Surrogate splits:
## reimbursement2008 < 79760 to the right, agree=0.87, adj=0.026, (0 split)
##
## Node number 924: 220 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.65 P(node) =0.0008005735
## class counts: 77 66 57 16 4
## probabilities: 0.350 0.300 0.259 0.073 0.018
## left son=1848 (132 obs) right son=1849 (88 obs)
## Primary splits:
## reimbursement2008 < 12810 to the right, improve=2.5787880, (0 missing)
## age < 76.5 to the left, improve=2.1718510, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.0384950, (0 missing)
## stroke < 0.5 to the left, improve=0.8392769, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7969697, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.636, adj=0.091, (0 split)
##
## Node number 925: 62 observations
## predicted class=B2 expected loss=0.5322581 P(node) =0.0002256162
## class counts: 11 29 14 7 1
## probabilities: 0.177 0.468 0.226 0.113 0.016
##
## Node number 944: 76 observations
## predicted class=B1 expected loss=0.4736842 P(node) =0.0002765618
## class counts: 40 18 12 5 1
## probabilities: 0.526 0.237 0.158 0.066 0.013
##
## Node number 945: 83 observations
## predicted class=B2 expected loss=0.6024096 P(node) =0.0003020345
## class counts: 25 33 21 3 1
## probabilities: 0.301 0.398 0.253 0.036 0.012
##
## Node number 948: 126 observations
## predicted class=B2 expected loss=0.5079365 P(node) =0.0004585103
## class counts: 22 62 33 9 0
## probabilities: 0.175 0.492 0.262 0.071 0.000
##
## Node number 949: 111 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6486486 P(node) =0.0004039257
## class counts: 13 39 39 17 3
## probabilities: 0.117 0.351 0.351 0.153 0.027
## left son=1898 (54 obs) right son=1899 (57 obs)
## Primary splits:
## age < 75.5 to the left, improve=1.9702330, (0 missing)
## reimbursement2008 < 23940 to the left, improve=1.5183950, (0 missing)
## stroke < 0.5 to the right, improve=1.4096100, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.1218360, (0 missing)
## kidney < 0.5 to the right, improve=0.9749865, (0 missing)
## Surrogate splits:
## reimbursement2008 < 36125 to the right, agree=0.586, adj=0.148, (0 split)
## depression < 0.5 to the right, agree=0.568, adj=0.111, (0 split)
## alzheimers < 0.5 to the right, agree=0.550, adj=0.074, (0 split)
## bucket2008 < 4.5 to the right, agree=0.532, adj=0.037, (0 split)
##
## Node number 962: 1636 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5177262 P(node) =0.005953356
## class counts: 789 562 198 80 7
## probabilities: 0.482 0.344 0.121 0.049 0.004
## left son=1924 (1127 obs) right son=1925 (509 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.275005, (0 missing)
## depression < 0.5 to the left, improve=2.200834, (0 missing)
## age < 56.5 to the right, improve=2.161392, (0 missing)
## reimbursement2008 < 3635 to the left, improve=1.571205, (0 missing)
## copd < 0.5 to the left, improve=1.483908, (0 missing)
## Surrogate splits:
## reimbursement2008 < 9210 to the left, agree=0.69, adj=0.004, (0 split)
##
## Node number 963: 578 observations, complexity param=0.0001439056
## predicted class=B2 expected loss=0.5916955 P(node) =0.002103325
## class counts: 220 236 92 27 3
## probabilities: 0.381 0.408 0.159 0.047 0.005
## left son=1926 (339 obs) right son=1927 (239 obs)
## Primary splits:
## depression < 0.5 to the left, improve=4.643194, (0 missing)
## copd < 0.5 to the left, improve=3.299852, (0 missing)
## reimbursement2008 < 4535 to the left, improve=1.771265, (0 missing)
## alzheimers < 0.5 to the left, improve=1.392171, (0 missing)
## age < 39.5 to the right, improve=1.271983, (0 missing)
## Surrogate splits:
## age < 43.5 to the right, agree=0.602, adj=0.038, (0 split)
##
## Node number 964: 1363 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5561262 P(node) =0.004959917
## class counts: 605 435 219 92 12
## probabilities: 0.444 0.319 0.161 0.067 0.009
## left son=1928 (798 obs) right son=1929 (565 obs)
## Primary splits:
## copd < 0.5 to the left, improve=7.963265, (0 missing)
## heart.failure < 0.5 to the left, improve=4.055475, (0 missing)
## reimbursement2008 < 29005 to the left, improve=3.506334, (0 missing)
## stroke < 0.5 to the left, improve=1.818093, (0 missing)
## alzheimers < 0.5 to the left, improve=1.815648, (0 missing)
## Surrogate splits:
## reimbursement2008 < 55265 to the left, agree=0.589, adj=0.009, (0 split)
## bucket2008 < 4.5 to the left, agree=0.589, adj=0.009, (0 split)
## age < 27.5 to the right, agree=0.588, adj=0.005, (0 split)
##
## Node number 965: 4200 observations, complexity param=0.0004649258
## predicted class=B2 expected loss=0.597619 P(node) =0.01528368
## class counts: 1619 1690 634 236 21
## probabilities: 0.385 0.402 0.151 0.056 0.005
## left son=1930 (1953 obs) right son=1931 (2247 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=7.153470, (0 missing)
## copd < 0.5 to the left, improve=3.796544, (0 missing)
## age < 49.5 to the left, improve=2.763159, (0 missing)
## reimbursement2008 < 3415 to the left, improve=2.562311, (0 missing)
## alzheimers < 0.5 to the left, improve=1.867356, (0 missing)
## Surrogate splits:
## reimbursement2008 < 4495 to the left, agree=0.563, adj=0.060, (0 split)
## copd < 0.5 to the left, agree=0.551, adj=0.035, (0 split)
## alzheimers < 0.5 to the left, agree=0.540, adj=0.010, (0 split)
## age < 45.5 to the left, agree=0.536, adj=0.002, (0 split)
##
## Node number 966: 2928 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5529372 P(node) =0.01065491
## class counts: 904 1309 506 190 19
## probabilities: 0.309 0.447 0.173 0.065 0.006
## left son=1932 (1987 obs) right son=1933 (941 obs)
## Primary splits:
## copd < 0.5 to the left, improve=5.088653, (0 missing)
## osteoporosis < 0.5 to the left, improve=4.205973, (0 missing)
## reimbursement2008 < 8045 to the left, improve=3.909055, (0 missing)
## heart.failure < 0.5 to the left, improve=3.356901, (0 missing)
## stroke < 0.5 to the left, improve=3.071040, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8235 to the left, agree=0.68, adj=0.005, (0 split)
##
## Node number 967: 1464 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6277322 P(node) =0.005327453
## class counts: 475 545 290 146 8
## probabilities: 0.324 0.372 0.198 0.100 0.005
## left son=1934 (36 obs) right son=1935 (1428 obs)
## Primary splits:
## reimbursement2008 < 8485 to the left, improve=3.191750, (0 missing)
## age < 78.5 to the left, improve=2.281932, (0 missing)
## heart.failure < 0.5 to the left, improve=2.180745, (0 missing)
## stroke < 0.5 to the left, improve=1.944689, (0 missing)
## copd < 0.5 to the left, improve=1.512341, (0 missing)
##
## Node number 976: 102 observations
## predicted class=B1 expected loss=0.4803922 P(node) =0.000371175
## class counts: 53 28 13 7 1
## probabilities: 0.520 0.275 0.127 0.069 0.010
##
## Node number 977: 961 observations
## predicted class=B2 expected loss=0.5848075 P(node) =0.003497051
## class counts: 283 399 179 88 12
## probabilities: 0.294 0.415 0.186 0.092 0.012
##
## Node number 992: 572 observations, complexity param=0.0001411382
## predicted class=B1 expected loss=0.5909091 P(node) =0.002081491
## class counts: 234 183 92 57 6
## probabilities: 0.409 0.320 0.161 0.100 0.010
## left son=1984 (101 obs) right son=1985 (471 obs)
## Primary splits:
## reimbursement2008 < 3545 to the left, improve=4.251847, (0 missing)
## bucket2008 < 2.5 to the left, improve=2.618377, (0 missing)
## age < 69.5 to the right, improve=2.566628, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.664728, (0 missing)
## heart.failure < 0.5 to the left, improve=1.532473, (0 missing)
##
## Node number 993: 392 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5790816 P(node) =0.001426476
## class counts: 122 165 75 25 5
## probabilities: 0.311 0.421 0.191 0.064 0.013
## left son=1986 (9 obs) right son=1987 (383 obs)
## Primary splits:
## reimbursement2008 < 14460 to the right, improve=2.744913, (0 missing)
## age < 48.5 to the left, improve=1.556717, (0 missing)
## copd < 0.5 to the left, improve=1.522824, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.319075, (0 missing)
## heart.failure < 0.5 to the left, improve=1.292043, (0 missing)
##
## Node number 994: 3172 observations
## predicted class=B2 expected loss=0.5630517 P(node) =0.01154281
## class counts: 709 1386 688 337 52
## probabilities: 0.224 0.437 0.217 0.106 0.016
##
## Node number 995: 3650 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6326027 P(node) =0.01328224
## class counts: 917 1341 812 510 70
## probabilities: 0.251 0.367 0.222 0.140 0.019
## left son=1990 (2424 obs) right son=1991 (1226 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=8.187206, (0 missing)
## depression < 0.5 to the left, improve=5.387571, (0 missing)
## copd < 0.5 to the left, improve=5.189054, (0 missing)
## alzheimers < 0.5 to the left, improve=3.710849, (0 missing)
## heart.failure < 0.5 to the left, improve=2.971629, (0 missing)
##
## Node number 1008: 973 observations
## predicted class=B2 expected loss=0.5611511 P(node) =0.003540718
## class counts: 160 427 184 174 28
## probabilities: 0.164 0.439 0.189 0.179 0.029
##
## Node number 1009: 318 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5283019 P(node) =0.001157193
## class counts: 23 150 96 45 4
## probabilities: 0.072 0.472 0.302 0.142 0.013
## left son=2018 (293 obs) right son=2019 (25 obs)
## Primary splits:
## reimbursement2008 < 16525 to the right, improve=7.797134, (0 missing)
## ihd < 0.5 to the left, improve=3.194748, (0 missing)
## heart.failure < 0.5 to the left, improve=1.170376, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.159119, (0 missing)
## age < 55 to the right, improve=1.113448, (0 missing)
##
## Node number 1014: 1286 observations
## predicted class=B2 expected loss=0.6251944 P(node) =0.004679716
## class counts: 74 482 303 348 79
## probabilities: 0.058 0.375 0.236 0.271 0.061
##
## Node number 1015: 251 observations, complexity param=6.088314e-05
## predicted class=B4 expected loss=0.6095618 P(node) =0.0009133816
## class counts: 13 77 50 98 13
## probabilities: 0.052 0.307 0.199 0.390 0.052
## left son=2030 (237 obs) right son=2031 (14 obs)
## Primary splits:
## reimbursement2008 < 101585 to the left, improve=3.425401, (0 missing)
## age < 61.5 to the left, improve=2.440583, (0 missing)
## heart.failure < 0.5 to the left, improve=2.158559, (0 missing)
## alzheimers < 0.5 to the left, improve=2.020557, (0 missing)
## cancer < 0.5 to the right, improve=1.778561, (0 missing)
##
## Node number 1016: 1317 observations, complexity param=0.0001439056
## predicted class=B2 expected loss=0.6757783 P(node) =0.004792524
## class counts: 269 427 234 313 74
## probabilities: 0.204 0.324 0.178 0.238 0.056
## left son=2032 (72 obs) right son=2033 (1245 obs)
## Primary splits:
## ihd < 0.5 to the left, improve=5.587185, (0 missing)
## reimbursement2008 < 22435 to the left, improve=5.039175, (0 missing)
## osteoporosis < 0.5 to the left, improve=3.150989, (0 missing)
## age < 50.5 to the right, improve=2.709964, (0 missing)
## heart.failure < 0.5 to the left, improve=2.161876, (0 missing)
##
## Node number 1017: 1172 observations, complexity param=0.0002036818
## predicted class=B4 expected loss=0.6953925 P(node) =0.004264873
## class counts: 272 265 202 357 76
## probabilities: 0.232 0.226 0.172 0.305 0.065
## left son=2034 (191 obs) right son=2035 (981 obs)
## Primary splits:
## reimbursement2008 < 43640 to the right, improve=6.105110, (0 missing)
## bucket2008 < 4.5 to the right, improve=4.295055, (0 missing)
## ihd < 0.5 to the left, improve=2.740224, (0 missing)
## heart.failure < 0.5 to the left, improve=2.395917, (0 missing)
## alzheimers < 0.5 to the right, improve=1.864237, (0 missing)
## Surrogate splits:
## bucket2008 < 4.5 to the right, agree=0.925, adj=0.539, (0 split)
##
## Node number 1652: 555 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5441441 P(node) =0.002019629
## class counts: 253 193 76 29 4
## probabilities: 0.456 0.348 0.137 0.052 0.007
## left son=3304 (265 obs) right son=3305 (290 obs)
## Primary splits:
## reimbursement2008 < 1955 to the left, improve=2.2492970, (0 missing)
## age < 44.5 to the right, improve=2.2010530, (0 missing)
## cancer < 0.5 to the left, improve=2.0271900, (0 missing)
## ihd < 0.5 to the left, improve=2.0227740, (0 missing)
## stroke < 0.5 to the left, improve=0.6965966, (0 missing)
## Surrogate splits:
## ihd < 0.5 to the left, agree=0.539, adj=0.034, (0 split)
## age < 90.5 to the right, agree=0.532, adj=0.019, (0 split)
## cancer < 0.5 to the right, agree=0.528, adj=0.011, (0 split)
##
## Node number 1653: 41 observations
## predicted class=B2 expected loss=0.4878049 P(node) =0.0001491978
## class counts: 13 21 1 5 1
## probabilities: 0.317 0.512 0.024 0.122 0.024
##
## Node number 1696: 263 observations
## predicted class=B1 expected loss=0.4980989 P(node) =0.0009570492
## class counts: 132 94 26 11 0
## probabilities: 0.502 0.357 0.099 0.042 0.000
##
## Node number 1697: 122 observations
## predicted class=B2 expected loss=0.5737705 P(node) =0.0004439544
## class counts: 43 52 13 12 2
## probabilities: 0.352 0.426 0.107 0.098 0.016
##
## Node number 1734: 171 observations
## predicted class=B1 expected loss=0.4736842 P(node) =0.0006222639
## class counts: 90 46 24 10 1
## probabilities: 0.526 0.269 0.140 0.058 0.006
##
## Node number 1735: 1057 observations, complexity param=6.918538e-05
## predicted class=B1 expected loss=0.5666982 P(node) =0.003846392
## class counts: 458 410 137 47 5
## probabilities: 0.433 0.388 0.130 0.044 0.005
## left son=3470 (840 obs) right son=3471 (217 obs)
## Primary splits:
## age < 83.5 to the left, improve=3.809819, (0 missing)
## kidney < 0.5 to the left, improve=2.564065, (0 missing)
## depression < 0.5 to the left, improve=1.351420, (0 missing)
## copd < 0.5 to the left, improve=1.145117, (0 missing)
## reimbursement2008 < 2975 to the left, improve=1.026292, (0 missing)
##
## Node number 1750: 1752 observations, complexity param=0.0001129105
## predicted class=B1 expected loss=0.6084475 P(node) =0.006375476
## class counts: 686 629 294 131 12
## probabilities: 0.392 0.359 0.168 0.075 0.007
## left son=3500 (1099 obs) right son=3501 (653 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.008256, (0 missing)
## age < 97.5 to the left, improve=2.167182, (0 missing)
## reimbursement2008 < 3055 to the left, improve=1.739250, (0 missing)
## alzheimers < 0.5 to the left, improve=1.536121, (0 missing)
## stroke < 0.5 to the left, improve=1.306511, (0 missing)
## Surrogate splits:
## age < 41.5 to the right, agree=0.628, adj=0.002, (0 split)
##
## Node number 1751: 39 observations
## predicted class=B2 expected loss=0.4615385 P(node) =0.0001419198
## class counts: 6 21 6 6 0
## probabilities: 0.154 0.538 0.154 0.154 0.000
##
## Node number 1752: 112 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0004075647
## class counts: 56 33 14 9 0
## probabilities: 0.500 0.295 0.125 0.080 0.000
##
## Node number 1753: 68 observations
## predicted class=B2 expected loss=0.4852941 P(node) =0.00024745
## class counts: 22 35 9 2 0
## probabilities: 0.324 0.515 0.132 0.029 0.000
##
## Node number 1754: 403 observations, complexity param=6.272808e-05
## predicted class=B1 expected loss=0.6253102 P(node) =0.001466505
## class counts: 151 147 79 20 6
## probabilities: 0.375 0.365 0.196 0.050 0.015
## left son=3508 (382 obs) right son=3509 (21 obs)
## Primary splits:
## reimbursement2008 < 2585 to the left, improve=1.3835870, (0 missing)
## age < 45.5 to the left, improve=1.1912610, (0 missing)
## depression < 0.5 to the left, improve=0.8974996, (0 missing)
## cancer < 0.5 to the right, improve=0.7248908, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2961750, (0 missing)
##
## Node number 1755: 30 observations
## predicted class=B2 expected loss=0.6333333 P(node) =0.0001091691
## class counts: 5 11 9 5 0
## probabilities: 0.167 0.367 0.300 0.167 0.000
##
## Node number 1766: 336 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5535714 P(node) =0.001222694
## class counts: 150 128 44 14 0
## probabilities: 0.446 0.381 0.131 0.042 0.000
## left son=3532 (322 obs) right son=3533 (14 obs)
## Primary splits:
## age < 90.5 to the left, improve=1.4565220, (0 missing)
## heart.failure < 0.5 to the left, improve=1.1063780, (0 missing)
## reimbursement2008 < 2325 to the right, improve=0.8683190, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8476676, (0 missing)
## stroke < 0.5 to the left, improve=0.4098214, (0 missing)
##
## Node number 1767: 13 observations
## predicted class=B2 expected loss=0.4615385 P(node) =4.730662e-05
## class counts: 1 7 2 2 1
## probabilities: 0.077 0.538 0.154 0.154 0.077
##
## Node number 1772: 1057 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.589404 P(node) =0.003846392
## class counts: 434 420 145 54 4
## probabilities: 0.411 0.397 0.137 0.051 0.004
## left son=3544 (1008 obs) right son=3545 (49 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=1.0144070, (0 missing)
## age < 46.5 to the left, improve=1.0088960, (0 missing)
## reimbursement2008 < 2535 to the right, improve=0.9481247, (0 missing)
## copd < 0.5 to the left, improve=0.6576908, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5672579, (0 missing)
##
## Node number 1773: 44 observations
## predicted class=B2 expected loss=0.4545455 P(node) =0.0001601147
## class counts: 11 24 5 3 1
## probabilities: 0.250 0.545 0.114 0.068 0.023
##
## Node number 1774: 403 observations
## predicted class=B2 expected loss=0.5756824 P(node) =0.001466505
## class counts: 130 171 70 28 4
## probabilities: 0.323 0.424 0.174 0.069 0.010
##
## Node number 1775: 29 observations
## predicted class=B1 expected loss=0.5517241 P(node) =0.0001055301
## class counts: 13 7 3 6 0
## probabilities: 0.448 0.241 0.103 0.207 0.000
##
## Node number 1802: 684 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5146199 P(node) =0.002489056
## class counts: 332 231 95 26 0
## probabilities: 0.485 0.338 0.139 0.038 0.000
## left son=3604 (286 obs) right son=3605 (398 obs)
## Primary splits:
## reimbursement2008 < 4365 to the left, improve=1.4254390, (0 missing)
## alzheimers < 0.5 to the left, improve=1.1902870, (0 missing)
## stroke < 0.5 to the left, improve=0.8341619, (0 missing)
## age < 34 to the left, improve=0.8175360, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3590913, (0 missing)
##
## Node number 1803: 9 observations
## predicted class=B3 expected loss=0.4444444 P(node) =3.275073e-05
## class counts: 3 1 5 0 0
## probabilities: 0.333 0.111 0.556 0.000 0.000
##
## Node number 1810: 1608 observations
## predicted class=B1 expected loss=0.5062189 P(node) =0.005851465
## class counts: 794 529 193 82 10
## probabilities: 0.494 0.329 0.120 0.051 0.006
##
## Node number 1811: 70 observations
## predicted class=B2 expected loss=0.4857143 P(node) =0.0002547279
## class counts: 26 36 6 2 0
## probabilities: 0.371 0.514 0.086 0.029 0.000
##
## Node number 1812: 405 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5012346 P(node) =0.001473783
## class counts: 202 140 42 18 3
## probabilities: 0.499 0.346 0.104 0.044 0.007
## left son=3624 (329 obs) right son=3625 (76 obs)
## Primary splits:
## age < 83.5 to the left, improve=1.8474760, (0 missing)
## reimbursement2008 < 14045 to the left, improve=1.4437850, (0 missing)
## kidney < 0.5 to the right, improve=1.0197570, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4573240, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4260458, (0 missing)
##
## Node number 1813: 452 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.579646 P(node) =0.001644815
## class counts: 190 173 58 30 1
## probabilities: 0.420 0.383 0.128 0.066 0.002
## left son=3626 (362 obs) right son=3627 (90 obs)
## Primary splits:
## reimbursement2008 < 3875 to the right, improve=2.645100, (0 missing)
## age < 84.5 to the left, improve=2.429876, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.612268, (0 missing)
## alzheimers < 0.5 to the right, improve=1.063100, (0 missing)
## kidney < 0.5 to the left, improve=0.663279, (0 missing)
## Surrogate splits:
## age < 32 to the right, agree=0.803, adj=0.011, (0 split)
##
## Node number 1820: 177 observations
## predicted class=B1 expected loss=0.559322 P(node) =0.0006440978
## class counts: 78 62 26 11 0
## probabilities: 0.441 0.350 0.147 0.062 0.000
##
## Node number 1821: 519 observations
## predicted class=B2 expected loss=0.6088632 P(node) =0.001888626
## class counts: 154 203 86 64 12
## probabilities: 0.297 0.391 0.166 0.123 0.023
##
## Node number 1826: 25 observations
## predicted class=B1 expected loss=0.32 P(node) =9.097426e-05
## class counts: 17 7 1 0 0
## probabilities: 0.680 0.280 0.040 0.000 0.000
##
## Node number 1827: 357 observations, complexity param=5.313437e-05
## predicted class=B2 expected loss=0.5826331 P(node) =0.001299112
## class counts: 140 149 48 19 1
## probabilities: 0.392 0.417 0.134 0.053 0.003
## left son=3654 (91 obs) right son=3655 (266 obs)
## Primary splits:
## age < 80.5 to the right, improve=2.1730570, (0 missing)
## reimbursement2008 < 4405 to the right, improve=2.0106540, (0 missing)
## depression < 0.5 to the left, improve=0.7793758, (0 missing)
## kidney < 0.5 to the left, improve=0.7738464, (0 missing)
## alzheimers < 0.5 to the right, improve=0.6298514, (0 missing)
##
## Node number 1846: 39 observations
## predicted class=B1 expected loss=0.4871795 P(node) =0.0001419198
## class counts: 20 12 4 3 0
## probabilities: 0.513 0.308 0.103 0.077 0.000
##
## Node number 1847: 254 observations
## predicted class=B2 expected loss=0.5787402 P(node) =0.0009242985
## class counts: 76 107 44 27 0
## probabilities: 0.299 0.421 0.173 0.106 0.000
##
## Node number 1848: 132 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5909091 P(node) =0.0004803441
## class counts: 54 42 26 9 1
## probabilities: 0.409 0.318 0.197 0.068 0.008
## left son=3696 (105 obs) right son=3697 (27 obs)
## Primary splits:
## age < 84.5 to the left, improve=4.2569990, (0 missing)
## reimbursement2008 < 13440 to the left, improve=1.5425260, (0 missing)
## ihd < 0.5 to the right, improve=1.3693110, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4363743, (0 missing)
## kidney < 0.5 to the left, improve=0.4353832, (0 missing)
##
## Node number 1849: 88 observations
## predicted class=B3 expected loss=0.6477273 P(node) =0.0003202294
## class counts: 23 24 31 7 3
## probabilities: 0.261 0.273 0.352 0.080 0.034
##
## Node number 1898: 54 observations
## predicted class=B3 expected loss=0.5555556 P(node) =0.0001965044
## class counts: 8 14 24 7 1
## probabilities: 0.148 0.259 0.444 0.130 0.019
##
## Node number 1899: 57 observations
## predicted class=B2 expected loss=0.5614035 P(node) =0.0002074213
## class counts: 5 25 15 10 2
## probabilities: 0.088 0.439 0.263 0.175 0.035
##
## Node number 1924: 1127 observations
## predicted class=B1 expected loss=0.4960071 P(node) =0.00410112
## class counts: 568 377 131 47 4
## probabilities: 0.504 0.335 0.116 0.042 0.004
##
## Node number 1925: 509 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5658153 P(node) =0.001852236
## class counts: 221 185 67 33 3
## probabilities: 0.434 0.363 0.132 0.065 0.006
## left son=3850 (137 obs) right son=3851 (372 obs)
## Primary splits:
## reimbursement2008 < 3775 to the left, improve=1.6880360, (0 missing)
## depression < 0.5 to the left, improve=1.6361880, (0 missing)
## age < 96.5 to the left, improve=1.5026800, (0 missing)
## copd < 0.5 to the left, improve=1.2566690, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7418596, (0 missing)
##
## Node number 1926: 339 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.5575221 P(node) =0.001233611
## class counts: 150 127 45 16 1
## probabilities: 0.442 0.375 0.133 0.047 0.003
## left son=3852 (211 obs) right son=3853 (128 obs)
## Primary splits:
## reimbursement2008 < 4905 to the left, improve=1.6963240, (0 missing)
## age < 45 to the right, improve=1.4829560, (0 missing)
## heart.failure < 0.5 to the left, improve=1.2573130, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6055009, (0 missing)
## copd < 0.5 to the left, improve=0.2708942, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.687, adj=0.172, (0 split)
## copd < 0.5 to the left, agree=0.664, adj=0.109, (0 split)
## stroke < 0.5 to the left, agree=0.652, adj=0.078, (0 split)
##
## Node number 1927: 239 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.5439331 P(node) =0.0008697139
## class counts: 70 109 47 11 2
## probabilities: 0.293 0.456 0.197 0.046 0.008
## left son=3854 (181 obs) right son=3855 (58 obs)
## Primary splits:
## copd < 0.5 to the left, improve=6.3468870, (0 missing)
## reimbursement2008 < 5790 to the right, improve=1.7891020, (0 missing)
## age < 60.5 to the left, improve=1.2691270, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8740764, (0 missing)
## stroke < 0.5 to the right, improve=0.6684821, (0 missing)
## Surrogate splits:
## age < 35 to the right, agree=0.762, adj=0.017, (0 split)
##
## Node number 1928: 798 observations
## predicted class=B1 expected loss=0.5075188 P(node) =0.002903898
## class counts: 393 256 95 51 3
## probabilities: 0.492 0.321 0.119 0.064 0.004
##
## Node number 1929: 565 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.6247788 P(node) =0.002056018
## class counts: 212 179 124 41 9
## probabilities: 0.375 0.317 0.219 0.073 0.016
## left son=3858 (116 obs) right son=3859 (449 obs)
## Primary splits:
## stroke < 0.5 to the right, improve=4.0381830, (0 missing)
## reimbursement2008 < 31655 to the left, improve=2.7523450, (0 missing)
## age < 42.5 to the right, improve=1.7655450, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.4692280, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5615109, (0 missing)
## Surrogate splits:
## reimbursement2008 < 61780 to the right, agree=0.8, adj=0.026, (0 split)
##
## Node number 1930: 1953 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.578085 P(node) =0.007106909
## class counts: 824 782 251 88 8
## probabilities: 0.422 0.400 0.129 0.045 0.004
## left son=3860 (343 obs) right son=3861 (1610 obs)
## Primary splits:
## reimbursement2008 < 3415 to the left, improve=3.4037160, (0 missing)
## age < 42.5 to the left, improve=3.2783080, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6509623, (0 missing)
## copd < 0.5 to the left, improve=0.5598170, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2946889, (0 missing)
##
## Node number 1931: 2247 observations, complexity param=0.0001605101
## predicted class=B2 expected loss=0.5959057 P(node) =0.008176767
## class counts: 795 908 383 148 13
## probabilities: 0.354 0.404 0.170 0.066 0.006
## left son=3862 (866 obs) right son=3863 (1381 obs)
## Primary splits:
## reimbursement2008 < 5335 to the right, improve=3.344298, (0 missing)
## copd < 0.5 to the left, improve=2.798571, (0 missing)
## age < 68.5 to the left, improve=2.255236, (0 missing)
## alzheimers < 0.5 to the left, improve=1.653597, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.448247, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.678, adj=0.165, (0 split)
## age < 34.5 to the left, agree=0.616, adj=0.005, (0 split)
##
## Node number 1932: 1987 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5546049 P(node) =0.007230634
## class counts: 662 885 320 111 9
## probabilities: 0.333 0.445 0.161 0.056 0.005
## left son=3864 (1964 obs) right son=3865 (23 obs)
## Primary splits:
## age < 98.5 to the left, improve=3.328502, (0 missing)
## heart.failure < 0.5 to the left, improve=3.141909, (0 missing)
## reimbursement2008 < 3085 to the left, improve=3.126917, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.906536, (0 missing)
## alzheimers < 0.5 to the left, improve=1.332632, (0 missing)
##
## Node number 1933: 941 observations
## predicted class=B2 expected loss=0.5494155 P(node) =0.003424271
## class counts: 242 424 186 79 10
## probabilities: 0.257 0.451 0.198 0.084 0.011
##
## Node number 1934: 36 observations
## predicted class=B1 expected loss=0.4444444 P(node) =0.0001310029
## class counts: 20 8 8 0 0
## probabilities: 0.556 0.222 0.222 0.000 0.000
##
## Node number 1935: 1428 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6239496 P(node) =0.00519645
## class counts: 455 537 282 146 8
## probabilities: 0.319 0.376 0.197 0.102 0.006
## left son=3870 (837 obs) right son=3871 (591 obs)
## Primary splits:
## age < 78.5 to the left, improve=2.474561, (0 missing)
## heart.failure < 0.5 to the left, improve=2.118405, (0 missing)
## stroke < 0.5 to the left, improve=1.930317, (0 missing)
## copd < 0.5 to the left, improve=1.447977, (0 missing)
## reimbursement2008 < 8670 to the right, improve=1.324274, (0 missing)
##
## Node number 1984: 101 observations
## predicted class=B2 expected loss=0.5247525 P(node) =0.000367536
## class counts: 33 48 15 4 1
## probabilities: 0.327 0.475 0.149 0.040 0.010
##
## Node number 1985: 471 observations, complexity param=5.811572e-05
## predicted class=B1 expected loss=0.5732484 P(node) =0.001713955
## class counts: 201 135 77 53 5
## probabilities: 0.427 0.287 0.163 0.113 0.011
## left son=3970 (346 obs) right son=3971 (125 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.506365, (0 missing)
## reimbursement2008 < 11515 to the left, improve=2.004779, (0 missing)
## heart.failure < 0.5 to the left, improve=1.922393, (0 missing)
## age < 72.5 to the right, improve=1.840715, (0 missing)
## bucket2008 < 2.5 to the left, improve=1.312872, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3600 to the right, agree=0.737, adj=0.008, (0 split)
##
## Node number 1986: 9 observations
## predicted class=B1 expected loss=0.2222222 P(node) =3.275073e-05
## class counts: 7 2 0 0 0
## probabilities: 0.778 0.222 0.000 0.000 0.000
##
## Node number 1987: 383 observations
## predicted class=B2 expected loss=0.5744125 P(node) =0.001393726
## class counts: 115 163 75 25 5
## probabilities: 0.300 0.426 0.196 0.065 0.013
##
## Node number 1990: 2424 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6575908 P(node) =0.008820864
## class counts: 663 830 547 335 49
## probabilities: 0.274 0.342 0.226 0.138 0.020
## left son=3980 (1234 obs) right son=3981 (1190 obs)
## Primary splits:
## depression < 0.5 to the left, improve=3.612677, (0 missing)
## age < 67.5 to the right, improve=3.329297, (0 missing)
## copd < 0.5 to the left, improve=3.109296, (0 missing)
## heart.failure < 0.5 to the left, improve=2.737555, (0 missing)
## reimbursement2008 < 9205 to the right, improve=2.610291, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.552, adj=0.087, (0 split)
## copd < 0.5 to the left, agree=0.540, adj=0.064, (0 split)
## age < 53.5 to the right, agree=0.526, adj=0.035, (0 split)
## reimbursement2008 < 12525 to the left, agree=0.522, adj=0.026, (0 split)
## heart.failure < 0.5 to the left, agree=0.520, adj=0.023, (0 split)
##
## Node number 1991: 1226 observations
## predicted class=B2 expected loss=0.5831974 P(node) =0.004461378
## class counts: 254 511 265 175 21
## probabilities: 0.207 0.417 0.216 0.143 0.017
##
## Node number 2018: 293 observations
## predicted class=B2 expected loss=0.4914676 P(node) =0.001066218
## class counts: 20 149 81 39 4
## probabilities: 0.068 0.509 0.276 0.133 0.014
##
## Node number 2019: 25 observations
## predicted class=B3 expected loss=0.4 P(node) =9.097426e-05
## class counts: 3 1 15 6 0
## probabilities: 0.120 0.040 0.600 0.240 0.000
##
## Node number 2030: 237 observations, complexity param=6.088314e-05
## predicted class=B4 expected loss=0.6329114 P(node) =0.000862436
## class counts: 13 76 49 87 12
## probabilities: 0.055 0.321 0.207 0.367 0.051
## left son=4060 (62 obs) right son=4061 (175 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=2.618202, (0 missing)
## reimbursement2008 < 90420 to the right, improve=2.488954, (0 missing)
## heart.failure < 0.5 to the left, improve=2.039633, (0 missing)
## age < 61.5 to the left, improve=1.881916, (0 missing)
## alzheimers < 0.5 to the left, improve=1.753135, (0 missing)
##
## Node number 2031: 14 observations
## predicted class=B4 expected loss=0.2142857 P(node) =5.094559e-05
## class counts: 0 1 1 11 1
## probabilities: 0.000 0.071 0.071 0.786 0.071
##
## Node number 2032: 72 observations
## predicted class=B1 expected loss=0.5694444 P(node) =0.0002620059
## class counts: 31 18 11 8 4
## probabilities: 0.431 0.250 0.153 0.111 0.056
##
## Node number 2033: 1245 observations
## predicted class=B2 expected loss=0.6714859 P(node) =0.004530518
## class counts: 238 409 223 305 70
## probabilities: 0.191 0.329 0.179 0.245 0.056
##
## Node number 2034: 191 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.6753927 P(node) =0.0006950434
## class counts: 29 62 44 42 14
## probabilities: 0.152 0.325 0.230 0.220 0.073
## left son=4068 (172 obs) right son=4069 (19 obs)
## Primary splits:
## age < 64.5 to the right, improve=2.5420870, (0 missing)
## reimbursement2008 < 71460 to the left, improve=2.3514440, (0 missing)
## alzheimers < 0.5 to the left, improve=1.2597430, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9221356, (0 missing)
## ihd < 0.5 to the right, improve=0.7384918, (0 missing)
##
## Node number 2035: 981 observations, complexity param=9.962695e-05
## predicted class=B4 expected loss=0.6788991 P(node) =0.00356983
## class counts: 243 203 158 315 62
## probabilities: 0.248 0.207 0.161 0.321 0.063
## left son=4070 (468 obs) right son=4071 (513 obs)
## Primary splits:
## reimbursement2008 < 23175 to the left, improve=5.196818, (0 missing)
## alzheimers < 0.5 to the right, improve=3.174409, (0 missing)
## ihd < 0.5 to the left, improve=2.640760, (0 missing)
## heart.failure < 0.5 to the left, improve=1.689264, (0 missing)
## age < 97.5 to the right, improve=1.688586, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.777, adj=0.532, (0 split)
## heart.failure < 0.5 to the left, agree=0.534, adj=0.024, (0 split)
## age < 53.5 to the left, agree=0.531, adj=0.017, (0 split)
## stroke < 0.5 to the right, agree=0.525, adj=0.004, (0 split)
##
## Node number 3304: 265 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.490566 P(node) =0.0009643272
## class counts: 135 82 34 13 1
## probabilities: 0.509 0.309 0.128 0.049 0.004
## left son=6608 (251 obs) right son=6609 (14 obs)
## Primary splits:
## stroke < 0.5 to the left, improve=3.807509, (0 missing)
## ihd < 0.5 to the left, improve=2.996787, (0 missing)
## cancer < 0.5 to the left, improve=2.863288, (0 missing)
## reimbursement2008 < 1815 to the left, improve=1.417998, (0 missing)
## depression < 0.5 to the left, improve=1.227469, (0 missing)
##
## Node number 3305: 290 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5931034 P(node) =0.001055301
## class counts: 118 111 42 16 3
## probabilities: 0.407 0.383 0.145 0.055 0.010
## left son=6610 (213 obs) right son=6611 (77 obs)
## Primary splits:
## age < 81.5 to the left, improve=1.9355560, (0 missing)
## reimbursement2008 < 2015 to the right, improve=1.1719950, (0 missing)
## ihd < 0.5 to the right, improve=0.8443893, (0 missing)
## stroke < 0.5 to the right, improve=0.5640543, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4757090, (0 missing)
##
## Node number 3470: 840 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5452381 P(node) =0.003056735
## class counts: 382 309 103 41 5
## probabilities: 0.455 0.368 0.123 0.049 0.006
## left son=6940 (71 obs) right son=6941 (769 obs)
## Primary splits:
## age < 54.5 to the left, improve=2.5793890, (0 missing)
## kidney < 0.5 to the left, improve=2.0319410, (0 missing)
## depression < 0.5 to the left, improve=1.3091120, (0 missing)
## reimbursement2008 < 2945 to the right, improve=1.1530320, (0 missing)
## copd < 0.5 to the left, improve=0.9762638, (0 missing)
##
## Node number 3471: 217 observations
## predicted class=B2 expected loss=0.5345622 P(node) =0.0007896566
## class counts: 76 101 34 6 0
## probabilities: 0.350 0.465 0.157 0.028 0.000
##
## Node number 3500: 1099 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.5814377 P(node) =0.003999229
## class counts: 460 388 168 76 7
## probabilities: 0.419 0.353 0.153 0.069 0.006
## left son=7000 (1074 obs) right son=7001 (25 obs)
## Primary splits:
## age < 95.5 to the left, improve=2.515661, (0 missing)
## copd < 0.5 to the left, improve=2.359857, (0 missing)
## cancer < 0.5 to the left, improve=1.641148, (0 missing)
## reimbursement2008 < 2575 to the right, improve=1.347245, (0 missing)
## stroke < 0.5 to the left, improve=1.145174, (0 missing)
##
## Node number 3501: 653 observations, complexity param=0.0001129105
## predicted class=B2 expected loss=0.6309342 P(node) =0.002376248
## class counts: 226 241 126 55 5
## probabilities: 0.346 0.369 0.193 0.084 0.008
## left son=7002 (303 obs) right son=7003 (350 obs)
## Primary splits:
## reimbursement2008 < 2655 to the left, improve=2.636734, (0 missing)
## cancer < 0.5 to the left, improve=1.461370, (0 missing)
## age < 55.5 to the right, improve=1.350106, (0 missing)
## alzheimers < 0.5 to the left, improve=1.189997, (0 missing)
## bucket2008 < 1.5 to the left, improve=1.091246, (0 missing)
## Surrogate splits:
## bucket2008 < 1.5 to the left, agree=0.548, adj=0.026, (0 split)
## copd < 0.5 to the right, agree=0.542, adj=0.013, (0 split)
## age < 47.5 to the left, agree=0.539, adj=0.007, (0 split)
##
## Node number 3508: 382 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.6230366 P(node) =0.001390087
## class counts: 142 144 74 18 4
## probabilities: 0.372 0.377 0.194 0.047 0.010
## left son=7016 (229 obs) right son=7017 (153 obs)
## Primary splits:
## depression < 0.5 to the left, improve=0.9679873, (0 missing)
## reimbursement2008 < 2275 to the left, improve=0.8225283, (0 missing)
## age < 75.5 to the right, improve=0.7274055, (0 missing)
## cancer < 0.5 to the right, improve=0.6895810, (0 missing)
## alzheimers < 0.5 to the right, improve=0.4333447, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.620, adj=0.052, (0 split)
## age < 50.5 to the right, agree=0.613, adj=0.033, (0 split)
##
## Node number 3509: 21 observations
## predicted class=B1 expected loss=0.5714286 P(node) =7.641838e-05
## class counts: 9 3 5 2 2
## probabilities: 0.429 0.143 0.238 0.095 0.095
##
## Node number 3532: 322 observations
## predicted class=B1 expected loss=0.5465839 P(node) =0.001171748
## class counts: 146 119 43 14 0
## probabilities: 0.453 0.370 0.134 0.043 0.000
##
## Node number 3533: 14 observations
## predicted class=B2 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 4 9 1 0 0
## probabilities: 0.286 0.643 0.071 0.000 0.000
##
## Node number 3544: 1008 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.5853175 P(node) =0.003668082
## class counts: 418 400 133 53 4
## probabilities: 0.415 0.397 0.132 0.053 0.004
## left son=7088 (275 obs) right son=7089 (733 obs)
## Primary splits:
## reimbursement2008 < 2535 to the right, improve=0.9732083, (0 missing)
## age < 39 to the left, improve=0.9699606, (0 missing)
## copd < 0.5 to the left, improve=0.8468269, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4615681, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4416739, (0 missing)
## Surrogate splits:
## age < 36.5 to the left, agree=0.728, adj=0.004, (0 split)
##
## Node number 3545: 49 observations
## predicted class=B2 expected loss=0.5918367 P(node) =0.0001783096
## class counts: 16 20 12 1 0
## probabilities: 0.327 0.408 0.245 0.020 0.000
##
## Node number 3604: 286 observations
## predicted class=B1 expected loss=0.4685315 P(node) =0.001040746
## class counts: 152 94 35 5 0
## probabilities: 0.531 0.329 0.122 0.017 0.000
##
## Node number 3605: 398 observations, complexity param=5.258089e-05
## predicted class=B1 expected loss=0.5477387 P(node) =0.00144831
## class counts: 180 137 60 21 0
## probabilities: 0.452 0.344 0.151 0.053 0.000
## left son=7210 (340 obs) right son=7211 (58 obs)
## Primary splits:
## reimbursement2008 < 4700 to the right, improve=5.7797840, (0 missing)
## alzheimers < 0.5 to the left, improve=1.1372610, (0 missing)
## age < 34.5 to the left, improve=0.9964329, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.5848011, (0 missing)
## kidney < 0.5 to the right, improve=0.4151452, (0 missing)
##
## Node number 3624: 329 observations
## predicted class=B1 expected loss=0.4832827 P(node) =0.001197221
## class counts: 170 105 35 16 3
## probabilities: 0.517 0.319 0.106 0.049 0.009
##
## Node number 3625: 76 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5394737 P(node) =0.0002765618
## class counts: 32 35 7 2 0
## probabilities: 0.421 0.461 0.092 0.026 0.000
## left son=7250 (21 obs) right son=7251 (55 obs)
## Primary splits:
## reimbursement2008 < 6785 to the right, improve=3.2066300, (0 missing)
## bucket2008 < 2.5 to the right, improve=3.1159910, (0 missing)
## alzheimers < 0.5 to the right, improve=1.9967220, (0 missing)
## age < 85.5 to the right, improve=1.1176690, (0 missing)
## kidney < 0.5 to the right, improve=0.9258269, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.921, adj=0.714, (0 split)
## kidney < 0.5 to the right, agree=0.789, adj=0.238, (0 split)
##
## Node number 3626: 362 observations
## predicted class=B1 expected loss=0.5552486 P(node) =0.001317307
## class counts: 161 128 47 25 1
## probabilities: 0.445 0.354 0.130 0.069 0.003
##
## Node number 3627: 90 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5 P(node) =0.0003275073
## class counts: 29 45 11 5 0
## probabilities: 0.322 0.500 0.122 0.056 0.000
## left son=7254 (21 obs) right son=7255 (69 obs)
## Primary splits:
## age < 69.5 to the left, improve=3.1544510, (0 missing)
## alzheimers < 0.5 to the right, improve=3.1535260, (0 missing)
## kidney < 0.5 to the left, improve=1.7000000, (0 missing)
## reimbursement2008 < 3185 to the left, improve=1.5133190, (0 missing)
## copd < 0.5 to the right, improve=0.3083333, (0 missing)
##
## Node number 3654: 91 observations
## predicted class=B1 expected loss=0.5054945 P(node) =0.0003311463
## class counts: 45 31 9 6 0
## probabilities: 0.495 0.341 0.099 0.066 0.000
##
## Node number 3655: 266 observations
## predicted class=B2 expected loss=0.556391 P(node) =0.0009679661
## class counts: 95 118 39 13 1
## probabilities: 0.357 0.444 0.147 0.049 0.004
##
## Node number 3696: 105 observations
## predicted class=B1 expected loss=0.5238095 P(node) =0.0003820919
## class counts: 50 35 16 3 1
## probabilities: 0.476 0.333 0.152 0.029 0.010
##
## Node number 3697: 27 observations
## predicted class=B3 expected loss=0.6296296 P(node) =9.82522e-05
## class counts: 4 7 10 6 0
## probabilities: 0.148 0.259 0.370 0.222 0.000
##
## Node number 3850: 137 observations
## predicted class=B1 expected loss=0.4963504 P(node) =0.000498539
## class counts: 69 41 17 9 1
## probabilities: 0.504 0.299 0.124 0.066 0.007
##
## Node number 3851: 372 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5913978 P(node) =0.001353697
## class counts: 152 144 50 24 2
## probabilities: 0.409 0.387 0.134 0.065 0.005
## left son=7702 (330 obs) right son=7703 (42 obs)
## Primary splits:
## reimbursement2008 < 4055 to the right, improve=3.5107950, (0 missing)
## age < 96 to the left, improve=2.0766610, (0 missing)
## depression < 0.5 to the left, improve=1.3295540, (0 missing)
## copd < 0.5 to the left, improve=1.2612770, (0 missing)
## heart.failure < 0.5 to the left, improve=0.1950857, (0 missing)
##
## Node number 3852: 211 observations, complexity param=9.409212e-05
## predicted class=B1 expected loss=0.563981 P(node) =0.0007678228
## class counts: 92 89 23 7 0
## probabilities: 0.436 0.422 0.109 0.033 0.000
## left son=7704 (142 obs) right son=7705 (69 obs)
## Primary splits:
## reimbursement2008 < 4075 to the left, improve=3.4884480, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9268848, (0 missing)
## heart.failure < 0.5 to the left, improve=0.8762896, (0 missing)
## age < 95 to the right, improve=0.6396384, (0 missing)
## stroke < 0.5 to the left, improve=0.3979516, (0 missing)
## Surrogate splits:
## age < 96.5 to the left, agree=0.687, adj=0.043, (0 split)
##
## Node number 3853: 128 observations
## predicted class=B1 expected loss=0.546875 P(node) =0.0004657882
## class counts: 58 38 22 9 1
## probabilities: 0.453 0.297 0.172 0.070 0.008
##
## Node number 3854: 181 observations
## predicted class=B2 expected loss=0.4861878 P(node) =0.0006586537
## class counts: 56 93 23 7 2
## probabilities: 0.309 0.514 0.127 0.039 0.011
##
## Node number 3855: 58 observations
## predicted class=B3 expected loss=0.5862069 P(node) =0.0002110603
## class counts: 14 16 24 4 0
## probabilities: 0.241 0.276 0.414 0.069 0.000
##
## Node number 3858: 116 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.5517241 P(node) =0.0004221206
## class counts: 41 52 14 7 2
## probabilities: 0.353 0.448 0.121 0.060 0.017
## left son=7716 (63 obs) right son=7717 (53 obs)
## Primary splits:
## age < 74.5 to the right, improve=2.8010500, (0 missing)
## reimbursement2008 < 17265 to the left, improve=2.4722090, (0 missing)
## bucket2008 < 4.5 to the left, improve=1.9776500, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5892240, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5845524, (0 missing)
## Surrogate splits:
## reimbursement2008 < 15590 to the right, agree=0.603, adj=0.132, (0 split)
## heart.failure < 0.5 to the right, agree=0.578, adj=0.075, (0 split)
## alzheimers < 0.5 to the right, agree=0.560, adj=0.038, (0 split)
## osteoporosis < 0.5 to the left, agree=0.552, adj=0.019, (0 split)
## bucket2008 < 3.5 to the right, agree=0.552, adj=0.019, (0 split)
##
## Node number 3859: 449 observations
## predicted class=B1 expected loss=0.6191537 P(node) =0.001633898
## class counts: 171 127 110 34 7
## probabilities: 0.381 0.283 0.245 0.076 0.016
##
## Node number 3860: 343 observations
## predicted class=B1 expected loss=0.5014577 P(node) =0.001248167
## class counts: 171 126 33 11 2
## probabilities: 0.499 0.367 0.096 0.032 0.006
##
## Node number 3861: 1610 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5925466 P(node) =0.005858742
## class counts: 653 656 218 77 6
## probabilities: 0.406 0.407 0.135 0.048 0.004
## left son=7722 (43 obs) right son=7723 (1567 obs)
## Primary splits:
## age < 42.5 to the left, improve=2.9787580, (0 missing)
## reimbursement2008 < 3475 to the right, improve=1.6291410, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4315089, (0 missing)
## copd < 0.5 to the left, improve=0.2703192, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2492479, (0 missing)
##
## Node number 3862: 866 observations, complexity param=0.0001605101
## predicted class=B1 expected loss=0.6120092 P(node) =0.003151348
## class counts: 336 322 139 64 5
## probabilities: 0.388 0.372 0.161 0.074 0.006
## left son=7724 (129 obs) right son=7725 (737 obs)
## Primary splits:
## reimbursement2008 < 8115 to the right, improve=2.0175360, (0 missing)
## age < 89.5 to the left, improve=1.8274460, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.6689370, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.3462440, (0 missing)
## stroke < 0.5 to the left, improve=0.5970317, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.984, adj=0.891, (0 split)
##
## Node number 3863: 1381 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.5756698 P(node) =0.005025418
## class counts: 459 586 244 84 8
## probabilities: 0.332 0.424 0.177 0.061 0.006
## left son=7726 (997 obs) right son=7727 (384 obs)
## Primary splits:
## copd < 0.5 to the left, improve=3.5627620, (0 missing)
## age < 37.5 to the right, improve=2.2016010, (0 missing)
## reimbursement2008 < 5195 to the left, improve=1.9417980, (0 missing)
## alzheimers < 0.5 to the left, improve=1.9283600, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3841814, (0 missing)
## Surrogate splits:
## age < 34 to the right, agree=0.723, adj=0.005, (0 split)
## reimbursement2008 < 5325 to the left, agree=0.723, adj=0.005, (0 split)
##
## Node number 3864: 1964 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.552444 P(node) =0.007146938
## class counts: 656 879 309 111 9
## probabilities: 0.334 0.448 0.157 0.057 0.005
## left son=7728 (22 obs) right son=7729 (1942 obs)
## Primary splits:
## reimbursement2008 < 3085 to the left, improve=3.418849, (0 missing)
## heart.failure < 0.5 to the left, improve=3.308540, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.919418, (0 missing)
## age < 66.5 to the right, improve=1.961336, (0 missing)
## alzheimers < 0.5 to the left, improve=1.448295, (0 missing)
##
## Node number 3865: 23 observations
## predicted class=B3 expected loss=0.5217391 P(node) =8.369632e-05
## class counts: 6 6 11 0 0
## probabilities: 0.261 0.261 0.478 0.000 0.000
##
## Node number 3870: 837 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6356033 P(node) =0.003045818
## class counts: 292 305 155 82 3
## probabilities: 0.349 0.364 0.185 0.098 0.004
## left son=7740 (639 obs) right son=7741 (198 obs)
## Primary splits:
## reimbursement2008 < 21320 to the left, improve=1.8992410, (0 missing)
## alzheimers < 0.5 to the left, improve=1.6748470, (0 missing)
## age < 49.5 to the left, improve=1.4981360, (0 missing)
## bucket2008 < 3.5 to the left, improve=1.4460210, (0 missing)
## stroke < 0.5 to the right, improve=0.7571134, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the left, agree=0.955, adj=0.808, (0 split)
##
## Node number 3871: 591 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.607445 P(node) =0.002150632
## class counts: 163 232 127 64 5
## probabilities: 0.276 0.393 0.215 0.108 0.008
## left son=7742 (122 obs) right son=7743 (469 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.4607800, (0 missing)
## reimbursement2008 < 8775 to the left, improve=1.9199300, (0 missing)
## stroke < 0.5 to the left, improve=1.2822570, (0 missing)
## copd < 0.5 to the left, improve=1.1325150, (0 missing)
## age < 80.5 to the right, improve=0.6677683, (0 missing)
##
## Node number 3970: 346 observations
## predicted class=B1 expected loss=0.5375723 P(node) =0.001259084
## class counts: 160 92 52 37 5
## probabilities: 0.462 0.266 0.150 0.107 0.014
##
## Node number 3971: 125 observations, complexity param=5.811572e-05
## predicted class=B2 expected loss=0.656 P(node) =0.0004548713
## class counts: 41 43 25 16 0
## probabilities: 0.328 0.344 0.200 0.128 0.000
## left son=7942 (106 obs) right son=7943 (19 obs)
## Primary splits:
## age < 62 to the right, improve=3.3415930, (0 missing)
## reimbursement2008 < 11475 to the left, improve=2.2730020, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7920000, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.7319750, (0 missing)
## stroke < 0.5 to the right, improve=0.5402967, (0 missing)
##
## Node number 3980: 1234 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6612642 P(node) =0.00449049
## class counts: 374 418 252 160 30
## probabilities: 0.303 0.339 0.204 0.130 0.024
## left son=7960 (349 obs) right son=7961 (885 obs)
## Primary splits:
## reimbursement2008 < 12135 to the right, improve=3.745241, (0 missing)
## age < 67.5 to the left, improve=3.421516, (0 missing)
## heart.failure < 0.5 to the left, improve=1.338981, (0 missing)
## bucket2008 < 2.5 to the right, improve=1.254047, (0 missing)
## copd < 0.5 to the right, improve=1.093433, (0 missing)
##
## Node number 3981: 1190 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6537815 P(node) =0.004330375
## class counts: 289 412 295 175 19
## probabilities: 0.243 0.346 0.248 0.147 0.016
## left son=7962 (547 obs) right son=7963 (643 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.655589, (0 missing)
## age < 82.5 to the right, improve=1.853724, (0 missing)
## stroke < 0.5 to the right, improve=1.583996, (0 missing)
## reimbursement2008 < 6355 to the right, improve=1.325969, (0 missing)
## heart.failure < 0.5 to the left, improve=1.085887, (0 missing)
## Surrogate splits:
## heart.failure < 0.5 to the left, agree=0.562, adj=0.048, (0 split)
## reimbursement2008 < 7315 to the left, agree=0.555, adj=0.031, (0 split)
## alzheimers < 0.5 to the left, agree=0.543, adj=0.005, (0 split)
## age < 28.5 to the left, agree=0.542, adj=0.004, (0 split)
##
## Node number 4060: 62 observations
## predicted class=B2 expected loss=0.5806452 P(node) =0.0002256162
## class counts: 3 26 17 15 1
## probabilities: 0.048 0.419 0.274 0.242 0.016
##
## Node number 4061: 175 observations
## predicted class=B4 expected loss=0.5885714 P(node) =0.0006368198
## class counts: 10 50 32 72 11
## probabilities: 0.057 0.286 0.183 0.411 0.063
##
## Node number 4068: 172 observations
## predicted class=B2 expected loss=0.6511628 P(node) =0.0006259029
## class counts: 27 60 35 39 11
## probabilities: 0.157 0.349 0.203 0.227 0.064
##
## Node number 4069: 19 observations
## predicted class=B3 expected loss=0.5263158 P(node) =6.914044e-05
## class counts: 2 2 9 3 3
## probabilities: 0.105 0.105 0.474 0.158 0.158
##
## Node number 4070: 468 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7200855 P(node) =0.001703038
## class counts: 131 112 77 122 26
## probabilities: 0.280 0.239 0.165 0.261 0.056
## left son=8140 (457 obs) right son=8141 (11 obs)
## Primary splits:
## age < 93.5 to the left, improve=1.955850, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.902288, (0 missing)
## reimbursement2008 < 19700 to the right, improve=1.873153, (0 missing)
## ihd < 0.5 to the right, improve=1.868954, (0 missing)
## alzheimers < 0.5 to the right, improve=1.716285, (0 missing)
##
## Node number 4071: 513 observations
## predicted class=B4 expected loss=0.6237817 P(node) =0.001866792
## class counts: 112 91 81 193 36
## probabilities: 0.218 0.177 0.158 0.376 0.070
##
## Node number 6608: 251 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.4701195 P(node) =0.0009133816
## class counts: 133 73 33 11 1
## probabilities: 0.530 0.291 0.131 0.044 0.004
## left son=13216 (235 obs) right son=13217 (16 obs)
## Primary splits:
## cancer < 0.5 to the left, improve=2.6866090, (0 missing)
## ihd < 0.5 to the left, improve=2.4584220, (0 missing)
## reimbursement2008 < 1815 to the left, improve=1.2636760, (0 missing)
## age < 60.5 to the right, improve=1.0616730, (0 missing)
## depression < 0.5 to the left, improve=0.7871996, (0 missing)
##
## Node number 6609: 14 observations
## predicted class=B2 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 2 9 1 2 0
## probabilities: 0.143 0.643 0.071 0.143 0.000
##
## Node number 6610: 213 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5633803 P(node) =0.0007751007
## class counts: 93 74 30 15 1
## probabilities: 0.437 0.347 0.141 0.070 0.005
## left son=13220 (201 obs) right son=13221 (12 obs)
## Primary splits:
## age < 44.5 to the right, improve=1.7572700, (0 missing)
## reimbursement2008 < 2005 to the right, improve=1.0657280, (0 missing)
## cancer < 0.5 to the right, improve=0.3892571, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3640621, (0 missing)
## depression < 0.5 to the right, improve=0.3467310, (0 missing)
##
## Node number 6611: 77 observations
## predicted class=B2 expected loss=0.5194805 P(node) =0.0002802007
## class counts: 25 37 12 1 2
## probabilities: 0.325 0.481 0.156 0.013 0.026
##
## Node number 6940: 71 observations
## predicted class=B1 expected loss=0.4366197 P(node) =0.0002583669
## class counts: 40 16 11 3 1
## probabilities: 0.563 0.225 0.155 0.042 0.014
##
## Node number 6941: 769 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5552666 P(node) =0.002798368
## class counts: 342 293 92 38 4
## probabilities: 0.445 0.381 0.120 0.049 0.005
## left son=13882 (472 obs) right son=13883 (297 obs)
## Primary splits:
## age < 70.5 to the right, improve=2.5248320, (0 missing)
## depression < 0.5 to the left, improve=1.8557100, (0 missing)
## kidney < 0.5 to the left, improve=1.7236880, (0 missing)
## reimbursement2008 < 2665 to the right, improve=1.1252400, (0 missing)
## copd < 0.5 to the left, improve=0.9387137, (0 missing)
##
## Node number 7000: 1074 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5772812 P(node) =0.003908254
## class counts: 454 373 166 74 7
## probabilities: 0.423 0.347 0.155 0.069 0.007
## left son=14000 (808 obs) right son=14001 (266 obs)
## Primary splits:
## copd < 0.5 to the left, improve=2.7468870, (0 missing)
## cancer < 0.5 to the left, improve=1.8315690, (0 missing)
## age < 78.5 to the left, improve=1.6074350, (0 missing)
## reimbursement2008 < 2575 to the right, improve=1.2651380, (0 missing)
## stroke < 0.5 to the left, improve=0.9951466, (0 missing)
##
## Node number 7001: 25 observations
## predicted class=B2 expected loss=0.4 P(node) =9.097426e-05
## class counts: 6 15 2 2 0
## probabilities: 0.240 0.600 0.080 0.080 0.000
##
## Node number 7002: 303 observations
## predicted class=B1 expected loss=0.6039604 P(node) =0.001102608
## class counts: 120 99 62 21 1
## probabilities: 0.396 0.327 0.205 0.069 0.003
##
## Node number 7003: 350 observations
## predicted class=B2 expected loss=0.5942857 P(node) =0.00127364
## class counts: 106 142 64 34 4
## probabilities: 0.303 0.406 0.183 0.097 0.011
##
## Node number 7016: 229 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5938865 P(node) =0.0008333242
## class counts: 93 82 44 8 2
## probabilities: 0.406 0.358 0.192 0.035 0.009
## left son=14032 (15 obs) right son=14033 (214 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=1.9222650, (0 missing)
## reimbursement2008 < 2515 to the left, improve=1.4164780, (0 missing)
## age < 94 to the left, improve=1.2547820, (0 missing)
## alzheimers < 0.5 to the right, improve=0.6631197, (0 missing)
## copd < 0.5 to the right, improve=0.2469242, (0 missing)
##
## Node number 7017: 153 observations, complexity param=6.272808e-05
## predicted class=B2 expected loss=0.5947712 P(node) =0.0005567625
## class counts: 49 62 30 10 2
## probabilities: 0.320 0.405 0.196 0.065 0.013
## left son=14034 (14 obs) right son=14035 (139 obs)
## Primary splits:
## reimbursement2008 < 2545 to the right, improve=2.7113570, (0 missing)
## age < 45 to the left, improve=1.6972360, (0 missing)
## cancer < 0.5 to the left, improve=0.6348039, (0 missing)
## copd < 0.5 to the right, improve=0.3887797, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2839287, (0 missing)
##
## Node number 7088: 275 observations, complexity param=7.748763e-05
## predicted class=B2 expected loss=0.56 P(node) =0.001000717
## class counts: 108 121 34 11 1
## probabilities: 0.393 0.440 0.124 0.040 0.004
## left son=14176 (44 obs) right son=14177 (231 obs)
## Primary splits:
## age < 63.5 to the left, improve=2.2068400, (0 missing)
## reimbursement2008 < 2555 to the right, improve=1.7374730, (0 missing)
## alzheimers < 0.5 to the right, improve=1.5968660, (0 missing)
## copd < 0.5 to the left, improve=0.9046397, (0 missing)
## heart.failure < 0.5 to the right, improve=0.5279104, (0 missing)
##
## Node number 7089: 733 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5770805 P(node) =0.002667365
## class counts: 310 279 99 42 3
## probabilities: 0.423 0.381 0.135 0.057 0.004
## left son=14178 (10 obs) right son=14179 (723 obs)
## Primary splits:
## age < 97.5 to the right, improve=1.1550490, (0 missing)
## heart.failure < 0.5 to the left, improve=1.1107930, (0 missing)
## reimbursement2008 < 2495 to the right, improve=0.7495829, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7242328, (0 missing)
## depression < 0.5 to the left, improve=0.5684301, (0 missing)
##
## Node number 7210: 340 observations
## predicted class=B1 expected loss=0.5088235 P(node) =0.00123725
## class counts: 167 107 48 18 0
## probabilities: 0.491 0.315 0.141 0.053 0.000
##
## Node number 7211: 58 observations
## predicted class=B2 expected loss=0.4827586 P(node) =0.0002110603
## class counts: 13 30 12 3 0
## probabilities: 0.224 0.517 0.207 0.052 0.000
##
## Node number 7250: 21 observations
## predicted class=B1 expected loss=0.3333333 P(node) =7.641838e-05
## class counts: 14 5 2 0 0
## probabilities: 0.667 0.238 0.095 0.000 0.000
##
## Node number 7251: 55 observations
## predicted class=B2 expected loss=0.4545455 P(node) =0.0002001434
## class counts: 18 30 5 2 0
## probabilities: 0.327 0.545 0.091 0.036 0.000
##
## Node number 7254: 21 observations
## predicted class=B1 expected loss=0.4285714 P(node) =7.641838e-05
## class counts: 12 6 1 2 0
## probabilities: 0.571 0.286 0.048 0.095 0.000
##
## Node number 7255: 69 observations
## predicted class=B2 expected loss=0.4347826 P(node) =0.000251089
## class counts: 17 39 10 3 0
## probabilities: 0.246 0.565 0.145 0.043 0.000
##
## Node number 7702: 330 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.569697 P(node) =0.00120086
## class counts: 142 119 44 23 2
## probabilities: 0.430 0.361 0.133 0.070 0.006
## left son=15404 (309 obs) right son=15405 (21 obs)
## Primary splits:
## reimbursement2008 < 4185 to the right, improve=2.2681710, (0 missing)
## age < 96 to the left, improve=2.1333520, (0 missing)
## depression < 0.5 to the left, improve=0.7533962, (0 missing)
## copd < 0.5 to the left, improve=0.6700147, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.3769465, (0 missing)
##
## Node number 7703: 42 observations
## predicted class=B2 expected loss=0.4047619 P(node) =0.0001528368
## class counts: 10 25 6 1 0
## probabilities: 0.238 0.595 0.143 0.024 0.000
##
## Node number 7704: 142 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0005167338
## class counts: 71 51 15 5 0
## probabilities: 0.500 0.359 0.106 0.035 0.000
##
## Node number 7705: 69 observations
## predicted class=B2 expected loss=0.4492754 P(node) =0.000251089
## class counts: 21 38 8 2 0
## probabilities: 0.304 0.551 0.116 0.029 0.000
##
## Node number 7716: 63 observations
## predicted class=B1 expected loss=0.5714286 P(node) =0.0002292551
## class counts: 27 21 10 4 1
## probabilities: 0.429 0.333 0.159 0.063 0.016
##
## Node number 7717: 53 observations
## predicted class=B2 expected loss=0.4150943 P(node) =0.0001928654
## class counts: 14 31 4 3 1
## probabilities: 0.264 0.585 0.075 0.057 0.019
##
## Node number 7722: 43 observations
## predicted class=B1 expected loss=0.3953488 P(node) =0.0001564757
## class counts: 26 11 3 3 0
## probabilities: 0.605 0.256 0.070 0.070 0.000
##
## Node number 7723: 1567 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5883854 P(node) =0.005702267
## class counts: 627 645 215 74 6
## probabilities: 0.400 0.412 0.137 0.047 0.004
## left son=15446 (1527 obs) right son=15447 (40 obs)
## Primary splits:
## age < 50.5 to the right, improve=1.7032880, (0 missing)
## reimbursement2008 < 3475 to the right, improve=1.5459000, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4552758, (0 missing)
## stroke < 0.5 to the left, improve=0.2471234, (0 missing)
## copd < 0.5 to the left, improve=0.2014160, (0 missing)
##
## Node number 7724: 129 observations
## predicted class=B2 expected loss=0.5271318 P(node) =0.0004694272
## class counts: 46 61 16 6 0
## probabilities: 0.357 0.473 0.124 0.047 0.000
##
## Node number 7725: 737 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.6065129 P(node) =0.002681921
## class counts: 290 261 123 58 5
## probabilities: 0.393 0.354 0.167 0.079 0.007
## left son=15450 (703 obs) right son=15451 (34 obs)
## Primary splits:
## age < 94.5 to the left, improve=1.9050170, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.5567680, (0 missing)
## reimbursement2008 < 6575 to the right, improve=1.5078350, (0 missing)
## copd < 0.5 to the left, improve=0.5423379, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5213862, (0 missing)
##
## Node number 7726: 997 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.5927783 P(node) =0.003628054
## class counts: 357 406 171 57 6
## probabilities: 0.358 0.407 0.172 0.057 0.006
## left son=15452 (297 obs) right son=15453 (700 obs)
## Primary splits:
## age < 69.5 to the left, improve=2.4458440, (0 missing)
## alzheimers < 0.5 to the left, improve=2.2624190, (0 missing)
## reimbursement2008 < 4135 to the right, improve=1.8635870, (0 missing)
## stroke < 0.5 to the right, improve=0.3191114, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3114115, (0 missing)
## Surrogate splits:
## reimbursement2008 < 5325 to the right, agree=0.703, adj=0.003, (0 split)
##
## Node number 7727: 384 observations
## predicted class=B2 expected loss=0.53125 P(node) =0.001397365
## class counts: 102 180 73 27 2
## probabilities: 0.266 0.469 0.190 0.070 0.005
##
## Node number 7728: 22 observations
## predicted class=B1 expected loss=0.3636364 P(node) =8.005735e-05
## class counts: 14 5 1 2 0
## probabilities: 0.636 0.227 0.045 0.091 0.000
##
## Node number 7729: 1942 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5499485 P(node) =0.007066881
## class counts: 642 874 308 109 9
## probabilities: 0.331 0.450 0.159 0.056 0.005
## left son=15458 (889 obs) right son=15459 (1053 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=3.215424, (0 missing)
## osteoporosis < 0.5 to the left, improve=2.693064, (0 missing)
## reimbursement2008 < 8025 to the left, improve=2.054172, (0 missing)
## age < 66.5 to the right, improve=1.953237, (0 missing)
## alzheimers < 0.5 to the left, improve=1.238773, (0 missing)
## Surrogate splits:
## reimbursement2008 < 3815 to the left, agree=0.545, adj=0.007, (0 split)
## age < 31.5 to the left, agree=0.544, adj=0.003, (0 split)
##
## Node number 7740: 639 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6353678 P(node) =0.002325302
## class counts: 233 221 124 59 2
## probabilities: 0.365 0.346 0.194 0.092 0.003
## left son=15480 (83 obs) right son=15481 (556 obs)
## Primary splits:
## age < 49.5 to the left, improve=1.7453130, (0 missing)
## stroke < 0.5 to the left, improve=1.0790130, (0 missing)
## reimbursement2008 < 10445 to the right, improve=1.0644190, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9731198, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.6917546, (0 missing)
##
## Node number 7741: 198 observations
## predicted class=B2 expected loss=0.5757576 P(node) =0.0007205162
## class counts: 59 84 31 23 1
## probabilities: 0.298 0.424 0.157 0.116 0.005
##
## Node number 7742: 122 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5983607 P(node) =0.0004439544
## class counts: 49 39 22 12 0
## probabilities: 0.402 0.320 0.180 0.098 0.000
## left son=15484 (40 obs) right son=15485 (82 obs)
## Primary splits:
## reimbursement2008 < 11560 to the left, improve=2.7817470, (0 missing)
## age < 80.5 to the right, improve=1.9103730, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4632510, (0 missing)
## copd < 0.5 to the left, improve=1.0641520, (0 missing)
## stroke < 0.5 to the left, improve=0.9890302, (0 missing)
##
## Node number 7743: 469 observations
## predicted class=B2 expected loss=0.5884861 P(node) =0.001706677
## class counts: 114 193 105 52 5
## probabilities: 0.243 0.412 0.224 0.111 0.011
##
## Node number 7942: 106 observations, complexity param=5.811572e-05
## predicted class=B1 expected loss=0.6320755 P(node) =0.0003857309
## class counts: 39 39 16 12 0
## probabilities: 0.368 0.368 0.151 0.113 0.000
## left son=15884 (93 obs) right son=15885 (13 obs)
## Primary splits:
## age < 67.5 to the right, improve=2.3273090, (0 missing)
## reimbursement2008 < 11575 to the left, improve=1.8244140, (0 missing)
## stroke < 0.5 to the right, improve=0.5034792, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4237564, (0 missing)
## heart.failure < 0.5 to the left, improve=0.3937905, (0 missing)
##
## Node number 7943: 19 observations
## predicted class=B3 expected loss=0.5263158 P(node) =6.914044e-05
## class counts: 2 4 9 4 0
## probabilities: 0.105 0.211 0.474 0.211 0.000
##
## Node number 7960: 349 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6475645 P(node) =0.001270001
## class counts: 123 103 55 57 11
## probabilities: 0.352 0.295 0.158 0.163 0.032
## left son=15920 (331 obs) right son=15921 (18 obs)
## Primary splits:
## age < 54 to the right, improve=2.0196730, (0 missing)
## alzheimers < 0.5 to the left, improve=1.9958000, (0 missing)
## reimbursement2008 < 15235 to the left, improve=1.7314800, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9381122, (0 missing)
## copd < 0.5 to the right, improve=0.4818854, (0 missing)
##
## Node number 7961: 885 observations
## predicted class=B2 expected loss=0.6440678 P(node) =0.003220489
## class counts: 251 315 197 103 19
## probabilities: 0.284 0.356 0.223 0.116 0.021
##
## Node number 7962: 547 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6709324 P(node) =0.001990517
## class counts: 154 180 136 65 12
## probabilities: 0.282 0.329 0.249 0.119 0.022
## left son=15924 (310 obs) right son=15925 (237 obs)
## Primary splits:
## reimbursement2008 < 9205 to the right, improve=2.7787810, (0 missing)
## age < 68.5 to the right, improve=2.5123800, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.9624740, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6055315, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5192530, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.835, adj=0.620, (0 split)
## age < 44.5 to the right, agree=0.581, adj=0.034, (0 split)
##
## Node number 7963: 643 observations
## predicted class=B2 expected loss=0.6391913 P(node) =0.002339858
## class counts: 135 232 159 110 7
## probabilities: 0.210 0.361 0.247 0.171 0.011
##
## Node number 8140: 457 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7133479 P(node) =0.00166301
## class counts: 131 107 73 120 26
## probabilities: 0.287 0.234 0.160 0.263 0.057
## left son=16280 (398 obs) right son=16281 (59 obs)
## Primary splits:
## age < 86.5 to the left, improve=2.218362, (0 missing)
## ihd < 0.5 to the right, improve=2.044330, (0 missing)
## reimbursement2008 < 19430 to the right, improve=1.853412, (0 missing)
## alzheimers < 0.5 to the right, improve=1.735004, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.352826, (0 missing)
##
## Node number 8141: 11 observations
## predicted class=B2 expected loss=0.5454545 P(node) =4.002868e-05
## class counts: 0 5 4 2 0
## probabilities: 0.000 0.455 0.364 0.182 0.000
##
## Node number 13216: 235 observations
## predicted class=B1 expected loss=0.4510638 P(node) =0.0008551581
## class counts: 129 64 30 11 1
## probabilities: 0.549 0.272 0.128 0.047 0.004
##
## Node number 13217: 16 observations
## predicted class=B2 expected loss=0.4375 P(node) =5.822353e-05
## class counts: 4 9 3 0 0
## probabilities: 0.250 0.562 0.188 0.000 0.000
##
## Node number 13220: 201 observations
## predicted class=B1 expected loss=0.5472637 P(node) =0.0007314331
## class counts: 91 67 29 14 0
## probabilities: 0.453 0.333 0.144 0.070 0.000
##
## Node number 13221: 12 observations
## predicted class=B2 expected loss=0.4166667 P(node) =4.366765e-05
## class counts: 2 7 1 1 1
## probabilities: 0.167 0.583 0.083 0.083 0.083
##
## Node number 13882: 472 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5275424 P(node) =0.001717594
## class counts: 223 163 57 25 4
## probabilities: 0.472 0.345 0.121 0.053 0.008
## left son=27764 (343 obs) right son=27765 (129 obs)
## Primary splits:
## age < 73.5 to the right, improve=4.630612, (0 missing)
## reimbursement2008 < 2805 to the right, improve=1.597068, (0 missing)
## depression < 0.5 to the left, improve=1.459900, (0 missing)
## kidney < 0.5 to the left, improve=1.335760, (0 missing)
## stroke < 0.5 to the left, improve=1.130037, (0 missing)
##
## Node number 13883: 297 observations, complexity param=6.088314e-05
## predicted class=B2 expected loss=0.5622896 P(node) =0.001080774
## class counts: 119 130 35 13 0
## probabilities: 0.401 0.438 0.118 0.044 0.000
## left son=27766 (218 obs) right son=27767 (79 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=1.3951400, (0 missing)
## reimbursement2008 < 2945 to the right, improve=1.0350230, (0 missing)
## depression < 0.5 to the left, improve=0.9259259, (0 missing)
## kidney < 0.5 to the left, improve=0.7583938, (0 missing)
## copd < 0.5 to the left, improve=0.3569379, (0 missing)
##
## Node number 14000: 808 observations
## predicted class=B1 expected loss=0.5569307 P(node) =0.002940288
## class counts: 358 273 111 61 5
## probabilities: 0.443 0.338 0.137 0.075 0.006
##
## Node number 14001: 266 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.6240602 P(node) =0.0009679661
## class counts: 96 100 55 13 2
## probabilities: 0.361 0.376 0.207 0.049 0.008
## left son=28002 (192 obs) right son=28003 (74 obs)
## Primary splits:
## reimbursement2008 < 2540 to the right, improve=2.9691060, (0 missing)
## age < 78.5 to the left, improve=2.6852920, (0 missing)
## cancer < 0.5 to the right, improve=2.3754980, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7018574, (0 missing)
## stroke < 0.5 to the left, improve=0.6157537, (0 missing)
## Surrogate splits:
## age < 50.5 to the right, agree=0.737, adj=0.054, (0 split)
##
## Node number 14032: 15 observations
## predicted class=B1 expected loss=0.3333333 P(node) =5.458456e-05
## class counts: 10 2 3 0 0
## probabilities: 0.667 0.133 0.200 0.000 0.000
##
## Node number 14033: 214 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6121495 P(node) =0.0007787397
## class counts: 83 80 41 8 2
## probabilities: 0.388 0.374 0.192 0.037 0.009
## left son=28066 (169 obs) right son=28067 (45 obs)
## Primary splits:
## reimbursement2008 < 2515 to the left, improve=1.6030020, (0 missing)
## age < 52.5 to the right, improve=0.9765448, (0 missing)
## alzheimers < 0.5 to the right, improve=0.7668533, (0 missing)
## copd < 0.5 to the right, improve=0.3681910, (0 missing)
## ihd < 0.5 to the right, improve=0.1207875, (0 missing)
##
## Node number 14034: 14 observations
## predicted class=B1 expected loss=0.3571429 P(node) =5.094559e-05
## class counts: 9 2 2 1 0
## probabilities: 0.643 0.143 0.143 0.071 0.000
##
## Node number 14035: 139 observations
## predicted class=B2 expected loss=0.5683453 P(node) =0.0005058169
## class counts: 40 60 28 9 2
## probabilities: 0.288 0.432 0.201 0.065 0.014
##
## Node number 14176: 44 observations
## predicted class=B2 expected loss=0.3863636 P(node) =0.0001601147
## class counts: 14 27 2 1 0
## probabilities: 0.318 0.614 0.045 0.023 0.000
##
## Node number 14177: 231 observations, complexity param=7.748763e-05
## predicted class=B1 expected loss=0.5930736 P(node) =0.0008406022
## class counts: 94 94 32 10 1
## probabilities: 0.407 0.407 0.139 0.043 0.004
## left son=28354 (169 obs) right son=28355 (62 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.4583990, (0 missing)
## reimbursement2008 < 2555 to the right, improve=1.0376560, (0 missing)
## age < 84.5 to the left, improve=1.0243680, (0 missing)
## copd < 0.5 to the left, improve=0.7240171, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4410819, (0 missing)
## Surrogate splits:
## age < 87.5 to the left, agree=0.745, adj=0.048, (0 split)
##
## Node number 14178: 10 observations
## predicted class=B1 expected loss=0.3 P(node) =3.63897e-05
## class counts: 7 2 1 0 0
## probabilities: 0.700 0.200 0.100 0.000 0.000
##
## Node number 14179: 723 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5809129 P(node) =0.002630976
## class counts: 303 277 98 42 3
## probabilities: 0.419 0.383 0.136 0.058 0.004
## left son=28358 (689 obs) right son=28359 (34 obs)
## Primary splits:
## age < 90.5 to the left, improve=1.6650270, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5078050, (0 missing)
## reimbursement2008 < 2495 to the right, improve=0.8133392, (0 missing)
## alzheimers < 0.5 to the left, improve=0.6699213, (0 missing)
## depression < 0.5 to the left, improve=0.5296598, (0 missing)
##
## Node number 15404: 309 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5728155 P(node) =0.001124442
## class counts: 132 117 38 20 2
## probabilities: 0.427 0.379 0.123 0.065 0.006
## left son=30808 (253 obs) right son=30809 (56 obs)
## Primary splits:
## reimbursement2008 < 4635 to the right, improve=2.0908250, (0 missing)
## age < 73.5 to the right, improve=1.8355900, (0 missing)
## depression < 0.5 to the left, improve=0.6554201, (0 missing)
## copd < 0.5 to the left, improve=0.3380891, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.2757170, (0 missing)
##
## Node number 15405: 21 observations
## predicted class=B1 expected loss=0.5238095 P(node) =7.641838e-05
## class counts: 10 2 6 3 0
## probabilities: 0.476 0.095 0.286 0.143 0.000
##
## Node number 15446: 1527 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5854617 P(node) =0.005556708
## class counts: 613 633 203 72 6
## probabilities: 0.401 0.415 0.133 0.047 0.004
## left son=30892 (1478 obs) right son=30893 (49 obs)
## Primary splits:
## reimbursement2008 < 3465 to the right, improve=1.7561930, (0 missing)
## age < 59.5 to the right, improve=1.2446620, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.3864080, (0 missing)
## stroke < 0.5 to the left, improve=0.3262151, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1237742, (0 missing)
##
## Node number 15447: 40 observations
## predicted class=B1 expected loss=0.65 P(node) =0.0001455588
## class counts: 14 12 12 2 0
## probabilities: 0.350 0.300 0.300 0.050 0.000
##
## Node number 15450: 703 observations, complexity param=8.855729e-05
## predicted class=B1 expected loss=0.598862 P(node) =0.002558196
## class counts: 282 244 119 53 5
## probabilities: 0.401 0.347 0.169 0.075 0.007
## left son=30900 (298 obs) right son=30901 (405 obs)
## Primary splits:
## reimbursement2008 < 6635 to the right, improve=1.9072210, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4867600, (0 missing)
## age < 74.5 to the left, improve=1.1374550, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4608058, (0 missing)
## alzheimers < 0.5 to the left, improve=0.4586126, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.596, adj=0.047, (0 split)
## age < 35.5 to the left, agree=0.578, adj=0.003, (0 split)
##
## Node number 15451: 34 observations
## predicted class=B2 expected loss=0.5 P(node) =0.000123725
## class counts: 8 17 4 5 0
## probabilities: 0.235 0.500 0.118 0.147 0.000
##
## Node number 15452: 297 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5757576 P(node) =0.001080774
## class counts: 126 113 45 12 1
## probabilities: 0.424 0.380 0.152 0.040 0.003
## left son=30904 (274 obs) right son=30905 (23 obs)
## Primary splits:
## reimbursement2008 < 5065 to the left, improve=2.3768610, (0 missing)
## alzheimers < 0.5 to the right, improve=2.2936150, (0 missing)
## age < 37.5 to the left, improve=1.9456180, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.7719678, (0 missing)
## stroke < 0.5 to the left, improve=0.6098027, (0 missing)
##
## Node number 15453: 700 observations
## predicted class=B2 expected loss=0.5814286 P(node) =0.002547279
## class counts: 231 293 126 45 5
## probabilities: 0.330 0.419 0.180 0.064 0.007
##
## Node number 15458: 889 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.003235045
## class counts: 327 381 133 45 3
## probabilities: 0.368 0.429 0.150 0.051 0.003
##
## Node number 15459: 1053 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5318139 P(node) =0.003831836
## class counts: 315 493 175 64 6
## probabilities: 0.299 0.468 0.166 0.061 0.006
## left son=30918 (721 obs) right son=30919 (332 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.319421, (0 missing)
## age < 65.5 to the right, improve=2.157808, (0 missing)
## reimbursement2008 < 4195 to the left, improve=2.005955, (0 missing)
## stroke < 0.5 to the left, improve=1.694776, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.685891, (0 missing)
##
## Node number 15480: 83 observations
## predicted class=B2 expected loss=0.5662651 P(node) =0.0003020345
## class counts: 29 36 8 10 0
## probabilities: 0.349 0.434 0.096 0.120 0.000
##
## Node number 15481: 556 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6330935 P(node) =0.002023268
## class counts: 204 185 116 49 2
## probabilities: 0.367 0.333 0.209 0.088 0.004
## left son=30962 (368 obs) right son=30963 (188 obs)
## Primary splits:
## age < 67.5 to the right, improve=1.7538220, (0 missing)
## reimbursement2008 < 17290 to the right, improve=1.5233210, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8892958, (0 missing)
## stroke < 0.5 to the left, improve=0.8663588, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8033839, (0 missing)
##
## Node number 15484: 40 observations
## predicted class=B1 expected loss=0.425 P(node) =0.0001455588
## class counts: 23 8 7 2 0
## probabilities: 0.575 0.200 0.175 0.050 0.000
##
## Node number 15485: 82 observations
## predicted class=B2 expected loss=0.6219512 P(node) =0.0002983956
## class counts: 26 31 15 10 0
## probabilities: 0.317 0.378 0.183 0.122 0.000
##
## Node number 15884: 93 observations, complexity param=5.811572e-05
## predicted class=B2 expected loss=0.5913978 P(node) =0.0003384243
## class counts: 32 38 15 8 0
## probabilities: 0.344 0.409 0.161 0.086 0.000
## left son=31768 (44 obs) right son=31769 (49 obs)
## Primary splits:
## reimbursement2008 < 6110 to the right, improve=2.8038180, (0 missing)
## age < 68.5 to the right, improve=0.9063337, (0 missing)
## heart.failure < 0.5 to the left, improve=0.4118188, (0 missing)
## alzheimers < 0.5 to the left, improve=0.3578690, (0 missing)
## stroke < 0.5 to the right, improve=0.3151562, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the right, agree=0.785, adj=0.545, (0 split)
## age < 73.5 to the left, agree=0.602, adj=0.159, (0 split)
## alzheimers < 0.5 to the right, agree=0.538, adj=0.023, (0 split)
##
## Node number 15885: 13 observations
## predicted class=B1 expected loss=0.4615385 P(node) =4.730662e-05
## class counts: 7 1 1 4 0
## probabilities: 0.538 0.077 0.077 0.308 0.000
##
## Node number 15920: 331 observations
## predicted class=B1 expected loss=0.6344411 P(node) =0.001204499
## class counts: 121 94 53 53 10
## probabilities: 0.366 0.284 0.160 0.160 0.030
##
## Node number 15921: 18 observations
## predicted class=B2 expected loss=0.5 P(node) =6.550147e-05
## class counts: 2 9 2 4 1
## probabilities: 0.111 0.500 0.111 0.222 0.056
##
## Node number 15924: 310 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6741935 P(node) =0.001128081
## class counts: 101 99 64 37 9
## probabilities: 0.326 0.319 0.206 0.119 0.029
## left son=31848 (50 obs) right son=31849 (260 obs)
## Primary splits:
## reimbursement2008 < 9955 to the left, improve=3.5194040, (0 missing)
## alzheimers < 0.5 to the left, improve=1.4052180, (0 missing)
## age < 60.5 to the right, improve=1.3545900, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9417634, (0 missing)
## stroke < 0.5 to the right, improve=0.4401818, (0 missing)
##
## Node number 15925: 237 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.6582278 P(node) =0.000862436
## class counts: 53 81 72 28 3
## probabilities: 0.224 0.342 0.304 0.118 0.013
## left son=31850 (56 obs) right son=31851 (181 obs)
## Primary splits:
## age < 67.5 to the left, improve=3.14488400, (0 missing)
## reimbursement2008 < 7130 to the left, improve=2.11196700, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.86604090, (0 missing)
## stroke < 0.5 to the left, improve=0.39390990, (0 missing)
## alzheimers < 0.5 to the right, improve=0.05008339, (0 missing)
##
## Node number 16280: 398 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.7236181 P(node) =0.00144831
## class counts: 110 101 63 99 25
## probabilities: 0.276 0.254 0.158 0.249 0.063
## left son=32560 (179 obs) right son=32561 (219 obs)
## Primary splits:
## alzheimers < 0.5 to the right, improve=2.797541, (0 missing)
## ihd < 0.5 to the right, improve=2.182276, (0 missing)
## reimbursement2008 < 15500 to the right, improve=1.710577, (0 missing)
## stroke < 0.5 to the right, improve=1.223226, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.211249, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the right, agree=0.563, adj=0.028, (0 split)
## reimbursement2008 < 15625 to the left, agree=0.563, adj=0.028, (0 split)
## age < 62.5 to the left, agree=0.555, adj=0.011, (0 split)
## osteoporosis < 0.5 to the right, agree=0.553, adj=0.006, (0 split)
##
## Node number 16281: 59 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6440678 P(node) =0.0002146993
## class counts: 21 6 10 21 1
## probabilities: 0.356 0.102 0.169 0.356 0.017
## left son=32562 (18 obs) right son=32563 (41 obs)
## Primary splits:
## reimbursement2008 < 19680 to the right, improve=1.9754260, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9501021, (0 missing)
## bucket2008 < 3.5 to the right, improve=0.8931654, (0 missing)
## age < 90.5 to the right, improve=0.7250257, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5260164, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.847, adj=0.5, (0 split)
##
## Node number 27764: 343 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.5568513 P(node) =0.001248167
## class counts: 152 136 37 16 2
## probabilities: 0.443 0.397 0.108 0.047 0.006
## left son=55528 (117 obs) right son=55529 (226 obs)
## Primary splits:
## reimbursement2008 < 2835 to the right, improve=1.9282960, (0 missing)
## stroke < 0.5 to the left, improve=1.1581140, (0 missing)
## age < 82.5 to the right, improve=1.0933820, (0 missing)
## kidney < 0.5 to the left, improve=1.0145490, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9380155, (0 missing)
##
## Node number 27765: 129 observations
## predicted class=B1 expected loss=0.4496124 P(node) =0.0004694272
## class counts: 71 27 20 9 2
## probabilities: 0.550 0.209 0.155 0.070 0.016
##
## Node number 27766: 218 observations, complexity param=6.088314e-05
## predicted class=B1 expected loss=0.5733945 P(node) =0.0007932956
## class counts: 93 89 28 8 0
## probabilities: 0.427 0.408 0.128 0.037 0.000
## left son=55532 (194 obs) right son=55533 (24 obs)
## Primary splits:
## reimbursement2008 < 2945 to the left, improve=1.9617420, (0 missing)
## depression < 0.5 to the left, improve=0.6526821, (0 missing)
## kidney < 0.5 to the left, improve=0.4610298, (0 missing)
## age < 57.5 to the left, improve=0.4574831, (0 missing)
## alzheimers < 0.5 to the left, improve=0.3559027, (0 missing)
##
## Node number 27767: 79 observations
## predicted class=B2 expected loss=0.4810127 P(node) =0.0002874787
## class counts: 26 41 7 5 0
## probabilities: 0.329 0.519 0.089 0.063 0.000
##
## Node number 28002: 192 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.59375 P(node) =0.0006986823
## class counts: 72 78 29 11 2
## probabilities: 0.375 0.406 0.151 0.057 0.010
## left son=56004 (124 obs) right son=56005 (68 obs)
## Primary splits:
## age < 78.5 to the left, improve=3.1968100, (0 missing)
## reimbursement2008 < 2885 to the left, improve=2.1236740, (0 missing)
## alzheimers < 0.5 to the right, improve=1.0053880, (0 missing)
## bucket2008 < 1.5 to the left, improve=0.7479369, (0 missing)
## stroke < 0.5 to the left, improve=0.5316513, (0 missing)
##
## Node number 28003: 74 observations, complexity param=7.19528e-05
## predicted class=B3 expected loss=0.6486486 P(node) =0.0002692838
## class counts: 24 22 26 2 0
## probabilities: 0.324 0.297 0.351 0.027 0.000
## left son=56006 (8 obs) right son=56007 (66 obs)
## Primary splits:
## cancer < 0.5 to the right, improve=6.4864860, (0 missing)
## age < 65 to the left, improve=1.7666590, (0 missing)
## alzheimers < 0.5 to the left, improve=1.7622440, (0 missing)
## reimbursement2008 < 2355 to the right, improve=1.0927360, (0 missing)
## stroke < 0.5 to the left, improve=0.8745462, (0 missing)
##
## Node number 28066: 169 observations
## predicted class=B1 expected loss=0.5798817 P(node) =0.000614986
## class counts: 71 58 32 6 2
## probabilities: 0.420 0.343 0.189 0.036 0.012
##
## Node number 28067: 45 observations
## predicted class=B2 expected loss=0.5111111 P(node) =0.0001637537
## class counts: 12 22 9 2 0
## probabilities: 0.267 0.489 0.200 0.044 0.000
##
## Node number 28354: 169 observations
## predicted class=B1 expected loss=0.5621302 P(node) =0.000614986
## class counts: 74 60 26 8 1
## probabilities: 0.438 0.355 0.154 0.047 0.006
##
## Node number 28355: 62 observations
## predicted class=B2 expected loss=0.4516129 P(node) =0.0002256162
## class counts: 20 34 6 2 0
## probabilities: 0.323 0.548 0.097 0.032 0.000
##
## Node number 28358: 689 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5718433 P(node) =0.002507251
## class counts: 295 261 92 38 3
## probabilities: 0.428 0.379 0.134 0.055 0.004
## left son=56716 (367 obs) right son=56717 (322 obs)
## Primary splits:
## heart.failure < 0.5 to the left, improve=1.5366830, (0 missing)
## reimbursement2008 < 2185 to the right, improve=0.9498001, (0 missing)
## age < 67.5 to the left, improve=0.9450906, (0 missing)
## copd < 0.5 to the left, improve=0.5052370, (0 missing)
## depression < 0.5 to the left, improve=0.4336301, (0 missing)
## Surrogate splits:
## copd < 0.5 to the left, agree=0.605, adj=0.155, (0 split)
## age < 85.5 to the left, agree=0.543, adj=0.022, (0 split)
## reimbursement2008 < 2515 to the left, agree=0.541, adj=0.019, (0 split)
## alzheimers < 0.5 to the left, agree=0.538, adj=0.012, (0 split)
##
## Node number 28359: 34 observations
## predicted class=B2 expected loss=0.5294118 P(node) =0.000123725
## class counts: 8 16 6 4 0
## probabilities: 0.235 0.471 0.176 0.118 0.000
##
## Node number 30808: 253 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5454545 P(node) =0.0009206595
## class counts: 115 89 30 17 2
## probabilities: 0.455 0.352 0.119 0.067 0.008
## left son=61616 (245 obs) right son=61617 (8 obs)
## Primary splits:
## age < 96 to the left, improve=1.6668230, (0 missing)
## reimbursement2008 < 8170 to the left, improve=1.5801570, (0 missing)
## depression < 0.5 to the left, improve=0.8012407, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6406559, (0 missing)
## stroke < 0.5 to the left, improve=0.4810539, (0 missing)
##
## Node number 30809: 56 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0002037823
## class counts: 17 28 8 3 0
## probabilities: 0.304 0.500 0.143 0.054 0.000
##
## Node number 30892: 1478 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5886333 P(node) =0.005378398
## class counts: 600 608 199 67 4
## probabilities: 0.406 0.411 0.135 0.045 0.003
## left son=61784 (759 obs) right son=61785 (719 obs)
## Primary splits:
## reimbursement2008 < 4655 to the left, improve=1.4912330, (0 missing)
## age < 59.5 to the right, improve=1.4379920, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.4252592, (0 missing)
## stroke < 0.5 to the left, improve=0.4189515, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.1287486, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.566, adj=0.108, (0 split)
## alzheimers < 0.5 to the left, agree=0.535, adj=0.043, (0 split)
## stroke < 0.5 to the left, agree=0.533, adj=0.040, (0 split)
## copd < 0.5 to the left, agree=0.530, adj=0.033, (0 split)
## age < 82.5 to the left, agree=0.526, adj=0.025, (0 split)
##
## Node number 30893: 49 observations
## predicted class=B2 expected loss=0.4897959 P(node) =0.0001783096
## class counts: 13 25 4 5 2
## probabilities: 0.265 0.510 0.082 0.102 0.041
##
## Node number 30900: 298 observations
## predicted class=B1 expected loss=0.5503356 P(node) =0.001084413
## class counts: 134 94 46 20 4
## probabilities: 0.450 0.315 0.154 0.067 0.013
##
## Node number 30901: 405 observations, complexity param=8.855729e-05
## predicted class=B2 expected loss=0.6296296 P(node) =0.001473783
## class counts: 148 150 73 33 1
## probabilities: 0.365 0.370 0.180 0.081 0.002
## left son=61802 (137 obs) right son=61803 (268 obs)
## Primary splits:
## reimbursement2008 < 5685 to the left, improve=1.4352860, (0 missing)
## age < 43 to the right, improve=1.2563810, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8108852, (0 missing)
## alzheimers < 0.5 to the right, improve=0.3644866, (0 missing)
## copd < 0.5 to the left, improve=0.3421456, (0 missing)
##
## Node number 30904: 274 observations, complexity param=7.19528e-05
## predicted class=B1 expected loss=0.5583942 P(node) =0.0009970779
## class counts: 121 99 42 11 1
## probabilities: 0.442 0.361 0.153 0.040 0.004
## left son=61808 (174 obs) right son=61809 (100 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.2349370, (0 missing)
## age < 37.5 to the left, improve=1.7714310, (0 missing)
## reimbursement2008 < 4990 to the right, improve=1.7636660, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8701440, (0 missing)
## stroke < 0.5 to the left, improve=0.4025273, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the left, agree=0.668, adj=0.09, (0 split)
## reimbursement2008 < 3085 to the right, agree=0.642, adj=0.02, (0 split)
##
## Node number 30905: 23 observations
## predicted class=B2 expected loss=0.3913043 P(node) =8.369632e-05
## class counts: 5 14 3 1 0
## probabilities: 0.217 0.609 0.130 0.043 0.000
##
## Node number 30918: 721 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5492372 P(node) =0.002623698
## class counts: 234 325 114 44 4
## probabilities: 0.325 0.451 0.158 0.061 0.006
## left son=61836 (109 obs) right son=61837 (612 obs)
## Primary splits:
## age < 86.5 to the right, improve=5.2024390, (0 missing)
## reimbursement2008 < 8105 to the left, improve=1.9497410, (0 missing)
## stroke < 0.5 to the left, improve=1.3441110, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.8592657, (0 missing)
## alzheimers < 0.5 to the left, improve=0.1415473, (0 missing)
##
## Node number 30919: 332 observations
## predicted class=B2 expected loss=0.4939759 P(node) =0.001208138
## class counts: 81 168 61 20 2
## probabilities: 0.244 0.506 0.184 0.060 0.006
##
## Node number 30962: 368 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.625 P(node) =0.001339141
## class counts: 138 132 66 32 0
## probabilities: 0.375 0.359 0.179 0.087 0.000
## left son=61924 (261 obs) right son=61925 (107 obs)
## Primary splits:
## reimbursement2008 < 10440 to the right, improve=2.0386870, (0 missing)
## age < 68.5 to the right, improve=2.0238320, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.0604210, (0 missing)
## alzheimers < 0.5 to the right, improve=0.8507150, (0 missing)
## heart.failure < 0.5 to the right, improve=0.3195541, (0 missing)
##
## Node number 30963: 188 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.6489362 P(node) =0.0006841264
## class counts: 66 53 50 17 2
## probabilities: 0.351 0.282 0.266 0.090 0.011
## left son=61926 (135 obs) right son=61927 (53 obs)
## Primary splits:
## age < 55.5 to the right, improve=1.3142350, (0 missing)
## reimbursement2008 < 8995 to the left, improve=1.1323620, (0 missing)
## stroke < 0.5 to the left, improve=0.7672950, (0 missing)
## heart.failure < 0.5 to the left, improve=0.7658279, (0 missing)
## bucket2008 < 3.5 to the right, improve=0.3998270, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8645 to the right, agree=0.723, adj=0.019, (0 split)
##
## Node number 31768: 44 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5227273 P(node) =0.0001601147
## class counts: 21 13 5 5 0
## probabilities: 0.477 0.295 0.114 0.114 0.000
## left son=63536 (26 obs) right son=63537 (18 obs)
## Primary splits:
## reimbursement2008 < 9180 to the left, improve=3.34188000, (0 missing)
## age < 73.5 to the right, improve=1.53473700, (0 missing)
## bucket2008 < 2.5 to the left, improve=1.08333300, (0 missing)
## alzheimers < 0.5 to the left, improve=0.99564270, (0 missing)
## copd < 0.5 to the right, improve=0.09090909, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.864, adj=0.667, (0 split)
## age < 82.5 to the left, agree=0.636, adj=0.111, (0 split)
## stroke < 0.5 to the left, agree=0.614, adj=0.056, (0 split)
##
## Node number 31769: 49 observations
## predicted class=B2 expected loss=0.4897959 P(node) =0.0001783096
## class counts: 11 25 10 3 0
## probabilities: 0.224 0.510 0.204 0.061 0.000
##
## Node number 31848: 50 observations
## predicted class=B2 expected loss=0.56 P(node) =0.0001819485
## class counts: 21 22 1 6 0
## probabilities: 0.420 0.440 0.020 0.120 0.000
##
## Node number 31849: 260 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6923077 P(node) =0.0009461323
## class counts: 80 77 63 31 9
## probabilities: 0.308 0.296 0.242 0.119 0.035
## left son=63698 (20 obs) right son=63699 (240 obs)
## Primary splits:
## reimbursement2008 < 14765 to the right, improve=1.866026, (0 missing)
## alzheimers < 0.5 to the left, improve=1.724179, (0 missing)
## age < 59 to the right, improve=1.389622, (0 missing)
## heart.failure < 0.5 to the left, improve=1.186623, (0 missing)
## stroke < 0.5 to the right, improve=0.396978, (0 missing)
##
## Node number 31850: 56 observations
## predicted class=B2 expected loss=0.5178571 P(node) =0.0002037823
## class counts: 11 27 9 9 0
## probabilities: 0.196 0.482 0.161 0.161 0.000
##
## Node number 31851: 181 observations, complexity param=5.534831e-05
## predicted class=B3 expected loss=0.6519337 P(node) =0.0006586537
## class counts: 42 54 63 19 3
## probabilities: 0.232 0.298 0.348 0.105 0.017
## left son=63702 (136 obs) right son=63703 (45 obs)
## Primary splits:
## reimbursement2008 < 6865 to the right, improve=2.6510090, (0 missing)
## age < 95.5 to the left, improve=1.1712710, (0 missing)
## bucket2008 < 2.5 to the right, improve=0.4758931, (0 missing)
## stroke < 0.5 to the left, improve=0.1841866, (0 missing)
## heart.failure < 0.5 to the left, improve=0.1010412, (0 missing)
##
## Node number 32560: 179 observations, complexity param=7.379774e-05
## predicted class=B1 expected loss=0.6815642 P(node) =0.0006513757
## class counts: 57 51 27 31 13
## probabilities: 0.318 0.285 0.151 0.173 0.073
## left son=65120 (38 obs) right son=65121 (141 obs)
## Primary splits:
## reimbursement2008 < 21440 to the right, improve=2.8400160, (0 missing)
## age < 70.5 to the left, improve=1.0471050, (0 missing)
## stroke < 0.5 to the right, improve=0.8887163, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8119666, (0 missing)
## heart.failure < 0.5 to the left, improve=0.6975642, (0 missing)
##
## Node number 32561: 219 observations
## predicted class=B4 expected loss=0.6894977 P(node) =0.0007969345
## class counts: 53 50 36 68 12
## probabilities: 0.242 0.228 0.164 0.311 0.055
##
## Node number 32562: 18 observations
## predicted class=B1 expected loss=0.4444444 P(node) =6.550147e-05
## class counts: 10 2 0 5 1
## probabilities: 0.556 0.111 0.000 0.278 0.056
##
## Node number 32563: 41 observations
## predicted class=B4 expected loss=0.6097561 P(node) =0.0001491978
## class counts: 11 4 10 16 0
## probabilities: 0.268 0.098 0.244 0.390 0.000
##
## Node number 55528: 117 observations, complexity param=5.165842e-05
## predicted class=B1 expected loss=0.4871795 P(node) =0.0004257595
## class counts: 60 38 15 3 1
## probabilities: 0.513 0.325 0.128 0.026 0.009
## left son=111056 (78 obs) right son=111057 (39 obs)
## Primary splits:
## reimbursement2008 < 2945 to the left, improve=2.9829060, (0 missing)
## kidney < 0.5 to the left, improve=1.2210830, (0 missing)
## age < 76.5 to the right, improve=1.1210830, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.9103070, (0 missing)
## stroke < 0.5 to the left, improve=0.2543679, (0 missing)
## Surrogate splits:
## age < 76.5 to the right, agree=0.684, adj=0.051, (0 split)
##
## Node number 55529: 226 observations, complexity param=5.165842e-05
## predicted class=B2 expected loss=0.5663717 P(node) =0.0008224073
## class counts: 92 98 22 13 1
## probabilities: 0.407 0.434 0.097 0.058 0.004
## left son=111058 (72 obs) right son=111059 (154 obs)
## Primary splits:
## age < 80.5 to the right, improve=1.5914710, (0 missing)
## stroke < 0.5 to the left, improve=1.3555880, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7668454, (0 missing)
## reimbursement2008 < 2795 to the left, improve=0.6874895, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.4980787, (0 missing)
##
## Node number 55532: 194 observations
## predicted class=B1 expected loss=0.556701 P(node) =0.0007059603
## class counts: 86 74 27 7 0
## probabilities: 0.443 0.381 0.139 0.036 0.000
##
## Node number 55533: 24 observations
## predicted class=B2 expected loss=0.375 P(node) =8.733529e-05
## class counts: 7 15 1 1 0
## probabilities: 0.292 0.625 0.042 0.042 0.000
##
## Node number 56004: 124 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5403226 P(node) =0.0004512323
## class counts: 57 47 16 4 0
## probabilities: 0.460 0.379 0.129 0.032 0.000
## left son=112008 (46 obs) right son=112009 (78 obs)
## Primary splits:
## age < 72.5 to the right, improve=2.8817380, (0 missing)
## reimbursement2008 < 2885 to the left, improve=1.5254660, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.4454760, (0 missing)
## alzheimers < 0.5 to the left, improve=0.7103829, (0 missing)
## stroke < 0.5 to the left, improve=0.4023915, (0 missing)
## Surrogate splits:
## stroke < 0.5 to the right, agree=0.661, adj=0.087, (0 split)
## reimbursement2008 < 2575 to the left, agree=0.645, adj=0.043, (0 split)
##
## Node number 56005: 68 observations
## predicted class=B2 expected loss=0.5441176 P(node) =0.00024745
## class counts: 15 31 13 7 2
## probabilities: 0.221 0.456 0.191 0.103 0.029
##
## Node number 56006: 8 observations
## predicted class=B2 expected loss=0 P(node) =2.911176e-05
## class counts: 0 8 0 0 0
## probabilities: 0.000 1.000 0.000 0.000 0.000
##
## Node number 56007: 66 observations, complexity param=5.534831e-05
## predicted class=B3 expected loss=0.6060606 P(node) =0.0002401721
## class counts: 24 14 26 2 0
## probabilities: 0.364 0.212 0.394 0.030 0.000
## left son=112014 (40 obs) right son=112015 (26 obs)
## Primary splits:
## alzheimers < 0.5 to the left, improve=2.3576920, (0 missing)
## age < 65 to the left, improve=1.8352940, (0 missing)
## reimbursement2008 < 2375 to the right, improve=1.1494920, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.1363636, (0 missing)
## Surrogate splits:
## age < 82.5 to the left, agree=0.652, adj=0.115, (0 split)
##
## Node number 56716: 367 observations
## predicted class=B1 expected loss=0.5395095 P(node) =0.001335502
## class counts: 169 131 51 13 3
## probabilities: 0.460 0.357 0.139 0.035 0.008
##
## Node number 56717: 322 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5962733 P(node) =0.001171748
## class counts: 126 130 41 25 0
## probabilities: 0.391 0.404 0.127 0.078 0.000
## left son=113434 (78 obs) right son=113435 (244 obs)
## Primary splits:
## age < 67.5 to the left, improve=2.0124890, (0 missing)
## reimbursement2008 < 2265 to the right, improve=1.1949400, (0 missing)
## alzheimers < 0.5 to the right, improve=0.3273471, (0 missing)
## depression < 0.5 to the right, improve=0.1786959, (0 missing)
## copd < 0.5 to the left, improve=0.1745923, (0 missing)
##
## Node number 61616: 245 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.5346939 P(node) =0.0008915478
## class counts: 114 87 27 16 1
## probabilities: 0.465 0.355 0.110 0.065 0.004
## left son=123232 (209 obs) right son=123233 (36 obs)
## Primary splits:
## reimbursement2008 < 8170 to the left, improve=1.7182870, (0 missing)
## age < 90.5 to the right, improve=1.6062760, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.7459219, (0 missing)
## depression < 0.5 to the left, improve=0.6596720, (0 missing)
## stroke < 0.5 to the left, improve=0.6366849, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.971, adj=0.806, (0 split)
##
## Node number 61617: 8 observations
## predicted class=B3 expected loss=0.625 P(node) =2.911176e-05
## class counts: 1 2 3 1 1
## probabilities: 0.125 0.250 0.375 0.125 0.125
##
## Node number 61784: 759 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5678524 P(node) =0.002761979
## class counts: 328 303 94 33 1
## probabilities: 0.432 0.399 0.124 0.043 0.001
## left son=123568 (158 obs) right son=123569 (601 obs)
## Primary splits:
## reimbursement2008 < 4315 to the right, improve=1.62186500, (0 missing)
## age < 82.5 to the right, improve=0.60286370, (0 missing)
## alzheimers < 0.5 to the right, improve=0.24697950, (0 missing)
## copd < 0.5 to the left, improve=0.10233690, (0 missing)
## stroke < 0.5 to the left, improve=0.09394217, (0 missing)
##
## Node number 61785: 719 observations, complexity param=9.962695e-05
## predicted class=B2 expected loss=0.5757997 P(node) =0.00261642
## class counts: 272 305 105 34 3
## probabilities: 0.378 0.424 0.146 0.047 0.004
## left son=123570 (346 obs) right son=123571 (373 obs)
## Primary splits:
## reimbursement2008 < 5835 to the left, improve=2.8015510, (0 missing)
## age < 59.5 to the right, improve=2.2849680, (0 missing)
## alzheimers < 0.5 to the right, improve=0.5855315, (0 missing)
## stroke < 0.5 to the right, improve=0.5109046, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.2469968, (0 missing)
## Surrogate splits:
## bucket2008 < 2.5 to the left, agree=0.590, adj=0.147, (0 split)
## alzheimers < 0.5 to the left, agree=0.537, adj=0.038, (0 split)
## age < 60.5 to the left, agree=0.527, adj=0.017, (0 split)
##
## Node number 61802: 137 observations
## predicted class=B1 expected loss=0.5839416 P(node) =0.000498539
## class counts: 57 43 22 15 0
## probabilities: 0.416 0.314 0.161 0.109 0.000
##
## Node number 61803: 268 observations
## predicted class=B2 expected loss=0.6007463 P(node) =0.0009752441
## class counts: 91 107 51 18 1
## probabilities: 0.340 0.399 0.190 0.067 0.004
##
## Node number 61808: 174 observations
## predicted class=B1 expected loss=0.5229885 P(node) =0.0006331809
## class counts: 83 53 29 8 1
## probabilities: 0.477 0.305 0.167 0.046 0.006
##
## Node number 61809: 100 observations, complexity param=7.19528e-05
## predicted class=B2 expected loss=0.54 P(node) =0.000363897
## class counts: 38 46 13 3 0
## probabilities: 0.380 0.460 0.130 0.030 0.000
## left son=123618 (26 obs) right son=123619 (74 obs)
## Primary splits:
## reimbursement2008 < 4355 to the right, improve=5.3372560, (0 missing)
## age < 62.5 to the right, improve=1.9704690, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.2703500, (0 missing)
## stroke < 0.5 to the left, improve=0.5886275, (0 missing)
## Surrogate splits:
## age < 50.5 to the left, agree=0.79, adj=0.192, (0 split)
##
## Node number 61836: 109 observations
## predicted class=B1 expected loss=0.5412844 P(node) =0.0003966478
## class counts: 50 33 16 9 1
## probabilities: 0.459 0.303 0.147 0.083 0.009
##
## Node number 61837: 612 observations
## predicted class=B2 expected loss=0.5228758 P(node) =0.00222705
## class counts: 184 292 98 35 3
## probabilities: 0.301 0.477 0.160 0.057 0.005
##
## Node number 61924: 261 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5862069 P(node) =0.0009497713
## class counts: 108 87 44 22 0
## probabilities: 0.414 0.333 0.169 0.084 0.000
## left son=123848 (92 obs) right son=123849 (169 obs)
## Primary splits:
## reimbursement2008 < 12585 to the left, improve=2.1315740, (0 missing)
## age < 77.5 to the right, improve=1.2761660, (0 missing)
## stroke < 0.5 to the left, improve=1.0543160, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.0296720, (0 missing)
## alzheimers < 0.5 to the left, improve=0.5202291, (0 missing)
##
## Node number 61925: 107 observations
## predicted class=B2 expected loss=0.5794393 P(node) =0.0003893698
## class counts: 30 45 22 10 0
## probabilities: 0.280 0.421 0.206 0.093 0.000
##
## Node number 61926: 135 observations
## predicted class=B1 expected loss=0.6074074 P(node) =0.000491261
## class counts: 53 34 36 12 0
## probabilities: 0.393 0.252 0.267 0.089 0.000
##
## Node number 61927: 53 observations
## predicted class=B2 expected loss=0.6415094 P(node) =0.0001928654
## class counts: 13 19 14 5 2
## probabilities: 0.245 0.358 0.264 0.094 0.038
##
## Node number 63536: 26 observations
## predicted class=B1 expected loss=0.3461538 P(node) =9.461323e-05
## class counts: 17 4 2 3 0
## probabilities: 0.654 0.154 0.077 0.115 0.000
##
## Node number 63537: 18 observations
## predicted class=B2 expected loss=0.5 P(node) =6.550147e-05
## class counts: 4 9 3 2 0
## probabilities: 0.222 0.500 0.167 0.111 0.000
##
## Node number 63698: 20 observations
## predicted class=B1 expected loss=0.45 P(node) =7.277941e-05
## class counts: 11 5 2 1 1
## probabilities: 0.550 0.250 0.100 0.050 0.050
##
## Node number 63699: 240 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.7 P(node) =0.0008733529
## class counts: 69 72 61 30 8
## probabilities: 0.288 0.300 0.254 0.125 0.033
## left son=127398 (201 obs) right son=127399 (39 obs)
## Primary splits:
## age < 61.5 to the right, improve=1.4580460, (0 missing)
## reimbursement2008 < 10970 to the right, improve=1.4206140, (0 missing)
## alzheimers < 0.5 to the left, improve=1.0755290, (0 missing)
## heart.failure < 0.5 to the left, improve=0.9752886, (0 missing)
## stroke < 0.5 to the right, improve=0.4524283, (0 missing)
##
## Node number 63702: 136 observations
## predicted class=B3 expected loss=0.6029412 P(node) =0.0004949
## class counts: 34 36 54 10 2
## probabilities: 0.250 0.265 0.397 0.074 0.015
##
## Node number 63703: 45 observations
## predicted class=B2 expected loss=0.6 P(node) =0.0001637537
## class counts: 8 18 9 9 1
## probabilities: 0.178 0.400 0.200 0.200 0.022
##
## Node number 65120: 38 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0001382809
## class counts: 9 19 4 5 1
## probabilities: 0.237 0.500 0.105 0.132 0.026
##
## Node number 65121: 141 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6595745 P(node) =0.0005130948
## class counts: 48 32 23 26 12
## probabilities: 0.340 0.227 0.163 0.184 0.085
## left son=130242 (89 obs) right son=130243 (52 obs)
## Primary splits:
## reimbursement2008 < 17585 to the right, improve=2.1889060, (0 missing)
## age < 47.5 to the right, improve=1.2186760, (0 missing)
## bucket2008 < 3.5 to the right, improve=1.1702130, (0 missing)
## stroke < 0.5 to the right, improve=0.9175166, (0 missing)
## heart.failure < 0.5 to the left, improve=0.5919705, (0 missing)
## Surrogate splits:
## bucket2008 < 3.5 to the right, agree=0.702, adj=0.192, (0 split)
## age < 47.5 to the right, agree=0.667, adj=0.096, (0 split)
##
## Node number 111056: 78 observations
## predicted class=B1 expected loss=0.4102564 P(node) =0.0002838397
## class counts: 46 19 11 2 0
## probabilities: 0.590 0.244 0.141 0.026 0.000
##
## Node number 111057: 39 observations
## predicted class=B2 expected loss=0.5128205 P(node) =0.0001419198
## class counts: 14 19 4 1 1
## probabilities: 0.359 0.487 0.103 0.026 0.026
##
## Node number 111058: 72 observations
## predicted class=B1 expected loss=0.4861111 P(node) =0.0002620059
## class counts: 37 29 5 1 0
## probabilities: 0.514 0.403 0.069 0.014 0.000
##
## Node number 111059: 154 observations
## predicted class=B2 expected loss=0.5519481 P(node) =0.0005604015
## class counts: 55 69 17 12 1
## probabilities: 0.357 0.448 0.110 0.078 0.006
##
## Node number 112008: 46 observations
## predicted class=B1 expected loss=0.4347826 P(node) =0.0001673926
## class counts: 26 10 8 2 0
## probabilities: 0.565 0.217 0.174 0.043 0.000
##
## Node number 112009: 78 observations
## predicted class=B2 expected loss=0.525641 P(node) =0.0002838397
## class counts: 31 37 8 2 0
## probabilities: 0.397 0.474 0.103 0.026 0.000
##
## Node number 112014: 40 observations
## predicted class=B1 expected loss=0.6 P(node) =0.0001455588
## class counts: 16 12 11 1 0
## probabilities: 0.400 0.300 0.275 0.025 0.000
##
## Node number 112015: 26 observations
## predicted class=B3 expected loss=0.4230769 P(node) =9.461323e-05
## class counts: 8 2 15 1 0
## probabilities: 0.308 0.077 0.577 0.038 0.000
##
## Node number 113434: 78 observations
## predicted class=B1 expected loss=0.5512821 P(node) =0.0002838397
## class counts: 35 23 15 5 0
## probabilities: 0.449 0.295 0.192 0.064 0.000
##
## Node number 113435: 244 observations
## predicted class=B2 expected loss=0.5614754 P(node) =0.0008879088
## class counts: 91 107 26 20 0
## probabilities: 0.373 0.439 0.107 0.082 0.000
##
## Node number 123232: 209 observations
## predicted class=B1 expected loss=0.507177 P(node) =0.0007605448
## class counts: 103 70 22 14 0
## probabilities: 0.493 0.335 0.105 0.067 0.000
##
## Node number 123233: 36 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5277778 P(node) =0.0001310029
## class counts: 11 17 5 2 1
## probabilities: 0.306 0.472 0.139 0.056 0.028
## left son=246466 (15 obs) right son=246467 (21 obs)
## Primary splits:
## age < 74.5 to the left, improve=5.3968250, (0 missing)
## reimbursement2008 < 8705 to the right, improve=1.5053320, (0 missing)
## copd < 0.5 to the right, improve=0.3703704, (0 missing)
## depression < 0.5 to the right, improve=0.3527778, (0 missing)
## heart.failure < 0.5 to the left, improve=0.2972583, (0 missing)
## Surrogate splits:
## reimbursement2008 < 8460 to the left, agree=0.611, adj=0.067, (0 split)
##
## Node number 123568: 158 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0005749573
## class counts: 79 55 15 8 1
## probabilities: 0.500 0.348 0.095 0.051 0.006
##
## Node number 123569: 601 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5856905 P(node) =0.002187021
## class counts: 249 248 79 25 0
## probabilities: 0.414 0.413 0.131 0.042 0.000
## left son=247138 (592 obs) right son=247139 (9 obs)
## Primary splits:
## reimbursement2008 < 4295 to the left, improve=3.0859230, (0 missing)
## age < 62.5 to the right, improve=0.8258296, (0 missing)
## copd < 0.5 to the left, improve=0.2730143, (0 missing)
## stroke < 0.5 to the left, improve=0.1209200, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1102521, (0 missing)
##
## Node number 123570: 346 observations
## predicted class=B2 expected loss=0.5202312 P(node) =0.001259084
## class counts: 122 166 44 13 1
## probabilities: 0.353 0.480 0.127 0.038 0.003
##
## Node number 123571: 373 observations, complexity param=9.962695e-05
## predicted class=B1 expected loss=0.5978552 P(node) =0.001357336
## class counts: 150 139 61 21 2
## probabilities: 0.402 0.373 0.164 0.056 0.005
## left son=247142 (124 obs) right son=247143 (249 obs)
## Primary splits:
## alzheimers < 0.5 to the right, improve=1.9370400, (0 missing)
## reimbursement2008 < 6045 to the right, improve=1.9317030, (0 missing)
## osteoporosis < 0.5 to the right, improve=1.1351660, (0 missing)
## age < 68.5 to the left, improve=0.9923350, (0 missing)
## stroke < 0.5 to the right, improve=0.8206414, (0 missing)
## Surrogate splits:
## age < 64.5 to the left, agree=0.673, adj=0.016, (0 split)
## stroke < 0.5 to the right, agree=0.673, adj=0.016, (0 split)
## reimbursement2008 < 5845 to the left, agree=0.670, adj=0.008, (0 split)
##
## Node number 123618: 26 observations
## predicted class=B1 expected loss=0.3846154 P(node) =9.461323e-05
## class counts: 16 4 4 2 0
## probabilities: 0.615 0.154 0.154 0.077 0.000
##
## Node number 123619: 74 observations
## predicted class=B2 expected loss=0.4324324 P(node) =0.0002692838
## class counts: 22 42 9 1 0
## probabilities: 0.297 0.568 0.122 0.014 0.000
##
## Node number 123848: 92 observations
## predicted class=B1 expected loss=0.5 P(node) =0.0003347853
## class counts: 46 23 17 6 0
## probabilities: 0.500 0.250 0.185 0.065 0.000
##
## Node number 123849: 169 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.6213018 P(node) =0.000614986
## class counts: 62 64 27 16 0
## probabilities: 0.367 0.379 0.160 0.095 0.000
## left son=247698 (109 obs) right son=247699 (60 obs)
## Primary splits:
## reimbursement2008 < 14485 to the right, improve=2.3703890, (0 missing)
## age < 77.5 to the right, improve=1.8205180, (0 missing)
## osteoporosis < 0.5 to the left, improve=1.5605270, (0 missing)
## alzheimers < 0.5 to the left, improve=0.9473954, (0 missing)
## stroke < 0.5 to the right, improve=0.8779250, (0 missing)
##
## Node number 127398: 201 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.6865672 P(node) =0.0007314331
## class counts: 63 58 49 27 4
## probabilities: 0.313 0.289 0.244 0.134 0.020
## left son=254796 (112 obs) right son=254797 (89 obs)
## Primary splits:
## reimbursement2008 < 12625 to the left, improve=2.6465710, (0 missing)
## age < 72.5 to the right, improve=1.5701210, (0 missing)
## heart.failure < 0.5 to the left, improve=1.5204340, (0 missing)
## alzheimers < 0.5 to the left, improve=0.8281641, (0 missing)
## stroke < 0.5 to the right, improve=0.4454147, (0 missing)
## Surrogate splits:
## age < 67.5 to the right, agree=0.587, adj=0.067, (0 split)
##
## Node number 127399: 39 observations
## predicted class=B2 expected loss=0.6410256 P(node) =0.0001419198
## class counts: 6 14 12 3 4
## probabilities: 0.154 0.359 0.308 0.077 0.103
##
## Node number 130242: 89 observations
## predicted class=B1 expected loss=0.5955056 P(node) =0.0003238684
## class counts: 36 15 17 14 7
## probabilities: 0.404 0.169 0.191 0.157 0.079
##
## Node number 130243: 52 observations
## predicted class=B2 expected loss=0.6730769 P(node) =0.0001892265
## class counts: 12 17 6 12 5
## probabilities: 0.231 0.327 0.115 0.231 0.096
##
## Node number 246466: 15 observations
## predicted class=B1 expected loss=0.4 P(node) =5.458456e-05
## class counts: 9 2 3 0 1
## probabilities: 0.600 0.133 0.200 0.000 0.067
##
## Node number 246467: 21 observations
## predicted class=B2 expected loss=0.2857143 P(node) =7.641838e-05
## class counts: 2 15 2 2 0
## probabilities: 0.095 0.714 0.095 0.095 0.000
##
## Node number 247138: 592 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5810811 P(node) =0.002154271
## class counts: 248 240 79 25 0
## probabilities: 0.419 0.405 0.133 0.042 0.000
## left son=494276 (135 obs) right son=494277 (457 obs)
## Primary splits:
## age < 82.5 to the right, improve=1.0162580, (0 missing)
## reimbursement2008 < 3485 to the left, improve=0.9533819, (0 missing)
## copd < 0.5 to the left, improve=0.2603666, (0 missing)
## alzheimers < 0.5 to the right, improve=0.1489946, (0 missing)
## stroke < 0.5 to the left, improve=0.1384892, (0 missing)
##
## Node number 247139: 9 observations
## predicted class=B2 expected loss=0.1111111 P(node) =3.275073e-05
## class counts: 1 8 0 0 0
## probabilities: 0.111 0.889 0.000 0.000 0.000
##
## Node number 247142: 124 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.516129 P(node) =0.0004512323
## class counts: 60 39 19 5 1
## probabilities: 0.484 0.315 0.153 0.040 0.008
## left son=494284 (114 obs) right son=494285 (10 obs)
## Primary splits:
## reimbursement2008 < 8555 to the left, improve=3.2894170, (0 missing)
## age < 62.5 to the right, improve=1.3134040, (0 missing)
## osteoporosis < 0.5 to the right, improve=0.8306452, (0 missing)
## stroke < 0.5 to the right, improve=0.6624062, (0 missing)
## bucket2008 < 2.5 to the left, improve=0.6169185, (0 missing)
##
## Node number 247143: 249 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5983936 P(node) =0.0009061036
## class counts: 90 100 42 16 1
## probabilities: 0.361 0.402 0.169 0.064 0.004
## left son=494286 (217 obs) right son=494287 (32 obs)
## Primary splits:
## reimbursement2008 < 6045 to the right, improve=2.8382200, (0 missing)
## age < 68.5 to the left, improve=1.5757780, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.8580882, (0 missing)
## copd < 0.5 to the left, improve=0.4427711, (0 missing)
## stroke < 0.5 to the right, improve=0.2244234, (0 missing)
##
## Node number 247698: 109 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5779817 P(node) =0.0003966478
## class counts: 46 34 19 10 0
## probabilities: 0.422 0.312 0.174 0.092 0.000
## left son=495396 (72 obs) right son=495397 (37 obs)
## Primary splits:
## osteoporosis < 0.5 to the left, improve=2.1824740, (0 missing)
## reimbursement2008 < 17235 to the right, improve=1.3957040, (0 missing)
## age < 70.5 to the left, improve=1.2827700, (0 missing)
## stroke < 0.5 to the left, improve=1.2406940, (0 missing)
## bucket2008 < 3.5 to the left, improve=0.2455781, (0 missing)
## Surrogate splits:
## reimbursement2008 < 14660 to the right, agree=0.679, adj=0.054, (0 split)
##
## Node number 247699: 60 observations
## predicted class=B2 expected loss=0.5 P(node) =0.0002183382
## class counts: 16 30 8 6 0
## probabilities: 0.267 0.500 0.133 0.100 0.000
##
## Node number 254796: 112 observations
## predicted class=B1 expected loss=0.6160714 P(node) =0.0004075647
## class counts: 43 27 31 10 1
## probabilities: 0.384 0.241 0.277 0.089 0.009
##
## Node number 254797: 89 observations
## predicted class=B2 expected loss=0.6516854 P(node) =0.0003238684
## class counts: 20 31 18 17 3
## probabilities: 0.225 0.348 0.202 0.191 0.034
##
## Node number 494276: 135 observations
## predicted class=B1 expected loss=0.5259259 P(node) =0.000491261
## class counts: 64 49 20 2 0
## probabilities: 0.474 0.363 0.148 0.015 0.000
##
## Node number 494277: 457 observations, complexity param=6.641797e-05
## predicted class=B2 expected loss=0.5820569 P(node) =0.00166301
## class counts: 184 191 59 23 0
## probabilities: 0.403 0.418 0.129 0.050 0.000
## left son=988554 (290 obs) right son=988555 (167 obs)
## Primary splits:
## age < 74.5 to the left, improve=0.9874503, (0 missing)
## reimbursement2008 < 3495 to the left, improve=0.9861916, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.3272674, (0 missing)
## stroke < 0.5 to the left, improve=0.2337493, (0 missing)
## copd < 0.5 to the left, improve=0.1994562, (0 missing)
## Surrogate splits:
## alzheimers < 0.5 to the left, agree=0.637, adj=0.006, (0 split)
##
## Node number 494284: 114 observations
## predicted class=B1 expected loss=0.4824561 P(node) =0.0004148426
## class counts: 59 32 18 4 1
## probabilities: 0.518 0.281 0.158 0.035 0.009
##
## Node number 494285: 10 observations
## predicted class=B2 expected loss=0.3 P(node) =3.63897e-05
## class counts: 1 7 1 1 0
## probabilities: 0.100 0.700 0.100 0.100 0.000
##
## Node number 494286: 217 observations
## predicted class=B2 expected loss=0.5714286 P(node) =0.0007896566
## class counts: 78 93 30 15 1
## probabilities: 0.359 0.429 0.138 0.069 0.005
##
## Node number 494287: 32 observations, complexity param=5.534831e-05
## predicted class=B1 expected loss=0.625 P(node) =0.0001164471
## class counts: 12 7 12 1 0
## probabilities: 0.375 0.219 0.375 0.031 0.000
## left son=988574 (11 obs) right son=988575 (21 obs)
## Primary splits:
## age < 72.5 to the left, improve=1.8097940, (0 missing)
## reimbursement2008 < 5975 to the left, improve=0.7232143, (0 missing)
## copd < 0.5 to the left, improve=0.6875000, (0 missing)
##
## Node number 495396: 72 observations
## predicted class=B1 expected loss=0.5138889 P(node) =0.0002620059
## class counts: 35 17 12 8 0
## probabilities: 0.486 0.236 0.167 0.111 0.000
##
## Node number 495397: 37 observations
## predicted class=B2 expected loss=0.5405405 P(node) =0.0001346419
## class counts: 11 17 7 2 0
## probabilities: 0.297 0.459 0.189 0.054 0.000
##
## Node number 988554: 290 observations, complexity param=6.641797e-05
## predicted class=B1 expected loss=0.5724138 P(node) =0.001055301
## class counts: 124 114 37 15 0
## probabilities: 0.428 0.393 0.128 0.052 0.000
## left son=1977108 (234 obs) right son=1977109 (56 obs)
## Primary splits:
## age < 62.5 to the right, improve=1.0825800, (0 missing)
## reimbursement2008 < 3945 to the right, improve=0.7040408, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.6026089, (0 missing)
## stroke < 0.5 to the left, improve=0.2655768, (0 missing)
## copd < 0.5 to the left, improve=0.1804923, (0 missing)
##
## Node number 988555: 167 observations, complexity param=5.534831e-05
## predicted class=B2 expected loss=0.5389222 P(node) =0.0006077081
## class counts: 60 77 22 8 0
## probabilities: 0.359 0.461 0.132 0.048 0.000
## left son=1977110 (39 obs) right son=1977111 (128 obs)
## Primary splits:
## reimbursement2008 < 4105 to the right, improve=1.3886510, (0 missing)
## osteoporosis < 0.5 to the left, improve=0.7439669, (0 missing)
## age < 81.5 to the left, improve=0.4824922, (0 missing)
## alzheimers < 0.5 to the right, improve=0.2060442, (0 missing)
## copd < 0.5 to the left, improve=0.1297289, (0 missing)
##
## Node number 988574: 11 observations
## predicted class=B1 expected loss=0.3636364 P(node) =4.002868e-05
## class counts: 7 2 2 0 0
## probabilities: 0.636 0.182 0.182 0.000 0.000
##
## Node number 988575: 21 observations
## predicted class=B3 expected loss=0.5238095 P(node) =7.641838e-05
## class counts: 5 5 10 1 0
## probabilities: 0.238 0.238 0.476 0.048 0.000
##
## Node number 1977108: 234 observations
## predicted class=B1 expected loss=0.5470085 P(node) =0.0008515191
## class counts: 106 89 28 11 0
## probabilities: 0.453 0.380 0.120 0.047 0.000
##
## Node number 1977109: 56 observations
## predicted class=B2 expected loss=0.5535714 P(node) =0.0002037823
## class counts: 18 25 9 4 0
## probabilities: 0.321 0.446 0.161 0.071 0.000
##
## Node number 1977110: 39 observations
## predicted class=B1 expected loss=0.5128205 P(node) =0.0001419198
## class counts: 19 14 5 1 0
## probabilities: 0.487 0.359 0.128 0.026 0.000
##
## Node number 1977111: 128 observations
## predicted class=B2 expected loss=0.5078125 P(node) =0.0004657882
## class counts: 41 63 17 7 0
## probabilities: 0.320 0.492 0.133 0.055 0.000
##
## n= 274803
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 274803 90337 B1 (0.67 0.19 0.089 0.043 0.0058)
## 2) reimbursement2008< 1565 165987 20938 B1 (0.87 0.074 0.037 0.014 0.0014) *
## 3) reimbursement2008>=1565 108816 68841 B2 (0.36 0.37 0.17 0.088 0.012)
## 6) reimbursement2008< 3065 39298 18853 B1 (0.52 0.31 0.12 0.045 0.0046)
## 12) reimbursement2008< 2175 20077 8527 B1 (0.58 0.27 0.11 0.042 0.0038)
## 24) diabetes< 0.5 8826 3280 B1 (0.63 0.24 0.091 0.035 0.0029) *
## 25) diabetes>=0.5 11251 5247 B1 (0.53 0.29 0.12 0.046 0.0045)
## 50) kidney< 0.5 9007 4045 B1 (0.55 0.28 0.12 0.042 0.0044)
## 100) reimbursement2008< 1875 4935 2071 B1 (0.58 0.26 0.11 0.042 0.0045) *
## 101) reimbursement2008>=1875 4072 1974 B1 (0.52 0.31 0.13 0.042 0.0044)
## 202) cancer< 0.5 3786 1807 B1 (0.52 0.3 0.13 0.041 0.0048) *
## 203) cancer>=0.5 286 167 B1 (0.42 0.4 0.13 0.056 0)
## 406) age< 73.5 128 64 B1 (0.5 0.33 0.11 0.062 0)
## 812) depression< 0.5 95 41 B1 (0.57 0.26 0.12 0.053 0) *
## 813) depression>=0.5 33 16 B2 (0.3 0.52 0.091 0.091 0) *
## 407) age>=73.5 158 85 B2 (0.35 0.46 0.14 0.051 0) *
## 51) kidney>=0.5 2244 1202 B1 (0.46 0.33 0.14 0.064 0.0049)
## 102) heart.failure< 0.5 992 477 B1 (0.52 0.29 0.13 0.057 0.002) *
## 103) heart.failure>=0.5 1252 725 B1 (0.42 0.36 0.15 0.069 0.0072)
## 206) arthritis< 0.5 904 486 B1 (0.46 0.34 0.13 0.063 0.0066)
## 412) reimbursement2008< 1735 270 125 B1 (0.54 0.27 0.13 0.056 0.0037) *
## 413) reimbursement2008>=1735 634 361 B1 (0.43 0.36 0.13 0.066 0.0079)
## 826) age< 91.5 596 330 B1 (0.45 0.36 0.13 0.057 0.0084)
## 1652) reimbursement2008>=1765 555 302 B1 (0.46 0.35 0.14 0.052 0.0072)
## 3304) reimbursement2008< 1955 265 130 B1 (0.51 0.31 0.13 0.049 0.0038)
## 6608) stroke< 0.5 251 118 B1 (0.53 0.29 0.13 0.044 0.004)
## 13216) cancer< 0.5 235 106 B1 (0.55 0.27 0.13 0.047 0.0043) *
## 13217) cancer>=0.5 16 7 B2 (0.25 0.56 0.19 0 0) *
## 6609) stroke>=0.5 14 5 B2 (0.14 0.64 0.071 0.14 0) *
## 3305) reimbursement2008>=1955 290 172 B1 (0.41 0.38 0.14 0.055 0.01)
## 6610) age< 81.5 213 120 B1 (0.44 0.35 0.14 0.07 0.0047)
## 13220) age>=44.5 201 110 B1 (0.45 0.33 0.14 0.07 0) *
## 13221) age< 44.5 12 5 B2 (0.17 0.58 0.083 0.083 0.083) *
## 6611) age>=81.5 77 40 B2 (0.32 0.48 0.16 0.013 0.026) *
## 1653) reimbursement2008< 1765 41 20 B2 (0.32 0.51 0.024 0.12 0.024) *
## 827) age>=91.5 38 21 B2 (0.18 0.45 0.16 0.21 0) *
## 207) arthritis>=0.5 348 205 B2 (0.31 0.41 0.18 0.086 0.0086) *
## 13) reimbursement2008>=2175 19221 10326 B1 (0.46 0.35 0.13 0.049 0.0054)
## 26) diabetes< 0.5 7137 3360 B1 (0.53 0.31 0.11 0.042 0.0046)
## 52) arthritis< 0.5 5554 2471 B1 (0.56 0.3 0.1 0.039 0.0049)
## 104) ihd< 0.5 2348 933 B1 (0.6 0.27 0.092 0.031 0.0051) *
## 105) ihd>=0.5 3206 1538 B1 (0.52 0.32 0.11 0.045 0.0047)
## 210) depression< 0.5 2325 1056 B1 (0.55 0.3 0.11 0.043 0.0052) *
## 211) depression>=0.5 881 482 B1 (0.45 0.36 0.13 0.052 0.0034)
## 422) kidney< 0.5 763 405 B1 (0.47 0.34 0.13 0.052 0.0039) *
## 423) kidney>=0.5 118 63 B2 (0.35 0.47 0.14 0.051 0)
## 846) reimbursement2008>=2865 22 10 B1 (0.55 0.23 0.18 0.045 0) *
## 847) reimbursement2008< 2865 96 46 B2 (0.3 0.52 0.12 0.052 0) *
## 53) arthritis>=0.5 1583 889 B1 (0.44 0.37 0.14 0.052 0.0038)
## 106) stroke< 0.5 1525 844 B1 (0.45 0.36 0.13 0.054 0.0039)
## 212) cancer< 0.5 1438 784 B1 (0.45 0.36 0.13 0.053 0.0042)
## 424) reimbursement2008>=2715 495 280 B1 (0.43 0.41 0.1 0.053 0.004)
## 848) reimbursement2008>=2795 385 210 B1 (0.45 0.38 0.1 0.06 0.0052)
## 1696) age< 80.5 263 131 B1 (0.5 0.36 0.099 0.042 0) *
## 1697) age>=80.5 122 70 B2 (0.35 0.43 0.11 0.098 0.016) *
## 849) reimbursement2008< 2795 110 54 B2 (0.36 0.51 0.1 0.027 0) *
## 425) reimbursement2008< 2715 943 504 B1 (0.47 0.33 0.15 0.053 0.0042) *
## 213) cancer>=0.5 87 48 B2 (0.31 0.45 0.17 0.069 0) *
## 107) stroke>=0.5 58 26 B2 (0.22 0.55 0.21 0.017 0) *
## 27) diabetes>=0.5 12084 6966 B1 (0.42 0.37 0.15 0.054 0.0059)
## 54) arthritis< 0.5 8413 4653 B1 (0.45 0.35 0.15 0.052 0.0056)
## 108) heart.failure< 0.5 4375 2220 B1 (0.49 0.34 0.13 0.039 0.0039)
## 216) cancer< 0.5 3992 1978 B1 (0.5 0.33 0.12 0.038 0.0033)
## 432) ihd< 0.5 1265 562 B1 (0.56 0.29 0.12 0.035 0.0032) *
## 433) ihd>=0.5 2727 1416 B1 (0.48 0.35 0.13 0.04 0.0033)
## 866) reimbursement2008< 2615 1499 736 B1 (0.51 0.33 0.13 0.035 0.002) *
## 867) reimbursement2008>=2615 1228 680 B1 (0.45 0.37 0.13 0.046 0.0049)
## 1734) reimbursement2008>=2995 171 81 B1 (0.53 0.27 0.14 0.058 0.0058) *
## 1735) reimbursement2008< 2995 1057 599 B1 (0.43 0.39 0.13 0.044 0.0047)
## 3470) age< 83.5 840 458 B1 (0.45 0.37 0.12 0.049 0.006)
## 6940) age< 54.5 71 31 B1 (0.56 0.23 0.15 0.042 0.014) *
## 6941) age>=54.5 769 427 B1 (0.44 0.38 0.12 0.049 0.0052)
## 13882) age>=70.5 472 249 B1 (0.47 0.35 0.12 0.053 0.0085)
## 27764) age>=73.5 343 191 B1 (0.44 0.4 0.11 0.047 0.0058)
## 55528) reimbursement2008>=2835 117 57 B1 (0.51 0.32 0.13 0.026 0.0085)
## 111056) reimbursement2008< 2945 78 32 B1 (0.59 0.24 0.14 0.026 0) *
## 111057) reimbursement2008>=2945 39 20 B2 (0.36 0.49 0.1 0.026 0.026) *
## 55529) reimbursement2008< 2835 226 128 B2 (0.41 0.43 0.097 0.058 0.0044)
## 111058) age>=80.5 72 35 B1 (0.51 0.4 0.069 0.014 0) *
## 111059) age< 80.5 154 85 B2 (0.36 0.45 0.11 0.078 0.0065) *
## 27765) age< 73.5 129 58 B1 (0.55 0.21 0.16 0.07 0.016) *
## 13883) age< 70.5 297 167 B2 (0.4 0.44 0.12 0.044 0)
## 27766) osteoporosis< 0.5 218 125 B1 (0.43 0.41 0.13 0.037 0)
## 55532) reimbursement2008< 2945 194 108 B1 (0.44 0.38 0.14 0.036 0) *
## 55533) reimbursement2008>=2945 24 9 B2 (0.29 0.62 0.042 0.042 0) *
## 27767) osteoporosis>=0.5 79 38 B2 (0.33 0.52 0.089 0.063 0) *
## 3471) age>=83.5 217 116 B2 (0.35 0.47 0.16 0.028 0) *
## 217) cancer>=0.5 383 220 B2 (0.37 0.43 0.15 0.044 0.01)
## 434) reimbursement2008< 2705 238 136 B1 (0.43 0.36 0.16 0.038 0.013)
## 868) depression< 0.5 167 84 B1 (0.5 0.3 0.15 0.042 0.012) *
## 869) depression>=0.5 71 35 B2 (0.27 0.51 0.18 0.028 0.014) *
## 435) reimbursement2008>=2705 145 68 B2 (0.27 0.53 0.14 0.055 0.0069) *
## 109) heart.failure>=0.5 4038 2433 B1 (0.4 0.36 0.17 0.066 0.0074)
## 218) kidney< 0.5 2819 1620 B1 (0.43 0.35 0.16 0.065 0.0064)
## 436) ihd< 0.5 635 319 B1 (0.5 0.31 0.15 0.041 0.0063) *
## 437) ihd>=0.5 2184 1301 B1 (0.4 0.36 0.16 0.072 0.0064)
## 874) reimbursement2008< 2315 393 202 B1 (0.49 0.34 0.12 0.051 0.0051) *
## 875) reimbursement2008>=2315 1791 1099 B1 (0.39 0.36 0.17 0.076 0.0067)
## 1750) age>=39.5 1752 1066 B1 (0.39 0.36 0.17 0.075 0.0068)
## 3500) depression< 0.5 1099 639 B1 (0.42 0.35 0.15 0.069 0.0064)
## 7000) age< 95.5 1074 620 B1 (0.42 0.35 0.15 0.069 0.0065)
## 14000) copd< 0.5 808 450 B1 (0.44 0.34 0.14 0.075 0.0062) *
## 14001) copd>=0.5 266 166 B2 (0.36 0.38 0.21 0.049 0.0075)
## 28002) reimbursement2008>=2540 192 114 B2 (0.38 0.41 0.15 0.057 0.01)
## 56004) age< 78.5 124 67 B1 (0.46 0.38 0.13 0.032 0)
## 112008) age>=72.5 46 20 B1 (0.57 0.22 0.17 0.043 0) *
## 112009) age< 72.5 78 41 B2 (0.4 0.47 0.1 0.026 0) *
## 56005) age>=78.5 68 37 B2 (0.22 0.46 0.19 0.1 0.029) *
## 28003) reimbursement2008< 2540 74 48 B3 (0.32 0.3 0.35 0.027 0)
## 56006) cancer>=0.5 8 0 B2 (0 1 0 0 0) *
## 56007) cancer< 0.5 66 40 B3 (0.36 0.21 0.39 0.03 0)
## 112014) alzheimers< 0.5 40 24 B1 (0.4 0.3 0.27 0.025 0) *
## 112015) alzheimers>=0.5 26 11 B3 (0.31 0.077 0.58 0.038 0) *
## 7001) age>=95.5 25 10 B2 (0.24 0.6 0.08 0.08 0) *
## 3501) depression>=0.5 653 412 B2 (0.35 0.37 0.19 0.084 0.0077)
## 7002) reimbursement2008< 2655 303 183 B1 (0.4 0.33 0.2 0.069 0.0033) *
## 7003) reimbursement2008>=2655 350 208 B2 (0.3 0.41 0.18 0.097 0.011) *
## 1751) age< 39.5 39 18 B2 (0.15 0.54 0.15 0.15 0) *
## 219) kidney>=0.5 1219 734 B2 (0.33 0.4 0.19 0.07 0.0098)
## 438) reimbursement2008< 2615 613 379 B1 (0.38 0.37 0.18 0.059 0.0098)
## 876) osteoporosis>=0.5 180 102 B1 (0.43 0.38 0.13 0.061 0)
## 1752) reimbursement2008< 2455 112 56 B1 (0.5 0.29 0.12 0.08 0) *
## 1753) reimbursement2008>=2455 68 33 B2 (0.32 0.51 0.13 0.029 0) *
## 877) osteoporosis< 0.5 433 275 B2 (0.36 0.36 0.2 0.058 0.014)
## 1754) stroke< 0.5 403 252 B1 (0.37 0.36 0.2 0.05 0.015)
## 3508) reimbursement2008< 2585 382 238 B2 (0.37 0.38 0.19 0.047 0.01)
## 7016) depression< 0.5 229 136 B1 (0.41 0.36 0.19 0.035 0.0087)
## 14032) cancer>=0.5 15 5 B1 (0.67 0.13 0.2 0 0) *
## 14033) cancer< 0.5 214 131 B1 (0.39 0.37 0.19 0.037 0.0093)
## 28066) reimbursement2008< 2515 169 98 B1 (0.42 0.34 0.19 0.036 0.012) *
## 28067) reimbursement2008>=2515 45 23 B2 (0.27 0.49 0.2 0.044 0) *
## 7017) depression>=0.5 153 91 B2 (0.32 0.41 0.2 0.065 0.013)
## 14034) reimbursement2008>=2545 14 5 B1 (0.64 0.14 0.14 0.071 0) *
## 14035) reimbursement2008< 2545 139 79 B2 (0.29 0.43 0.2 0.065 0.014) *
## 3509) reimbursement2008>=2585 21 12 B1 (0.43 0.14 0.24 0.095 0.095) *
## 1755) stroke>=0.5 30 19 B2 (0.17 0.37 0.3 0.17 0) *
## 439) reimbursement2008>=2615 606 347 B2 (0.28 0.43 0.2 0.081 0.0099) *
## 55) arthritis>=0.5 3671 2129 B2 (0.37 0.42 0.15 0.057 0.0065)
## 110) reimbursement2008< 2665 2068 1224 B1 (0.41 0.4 0.14 0.057 0.0048)
## 220) ihd< 0.5 517 274 B1 (0.47 0.37 0.11 0.048 0.0019)
## 440) reimbursement2008< 2295 143 57 B1 (0.6 0.26 0.077 0.063 0) *
## 441) reimbursement2008>=2295 374 217 B1 (0.42 0.41 0.12 0.043 0.0027)
## 882) reimbursement2008< 2315 25 6 B2 (0.24 0.76 0 0 0) *
## 883) reimbursement2008>=2315 349 198 B1 (0.43 0.39 0.13 0.046 0.0029)
## 1766) cancer< 0.5 336 186 B1 (0.45 0.38 0.13 0.042 0)
## 3532) age< 90.5 322 176 B1 (0.45 0.37 0.13 0.043 0) *
## 3533) age>=90.5 14 5 B2 (0.29 0.64 0.071 0 0) *
## 1767) cancer>=0.5 13 6 B2 (0.077 0.54 0.15 0.15 0.077) *
## 221) ihd>=0.5 1551 925 B2 (0.39 0.4 0.14 0.059 0.0058)
## 442) age< 35 18 5 B1 (0.72 0.22 0 0.056 0) *
## 443) age>=35 1533 911 B2 (0.38 0.41 0.15 0.059 0.0059)
## 886) kidney< 0.5 1101 656 B1 (0.4 0.4 0.14 0.052 0.0045)
## 1772) stroke< 0.5 1057 623 B1 (0.41 0.4 0.14 0.051 0.0038)
## 3544) cancer< 0.5 1008 590 B1 (0.41 0.4 0.13 0.053 0.004)
## 7088) reimbursement2008>=2535 275 154 B2 (0.39 0.44 0.12 0.04 0.0036)
## 14176) age< 63.5 44 17 B2 (0.32 0.61 0.045 0.023 0) *
## 14177) age>=63.5 231 137 B1 (0.41 0.41 0.14 0.043 0.0043)
## 28354) alzheimers< 0.5 169 95 B1 (0.44 0.36 0.15 0.047 0.0059) *
## 28355) alzheimers>=0.5 62 28 B2 (0.32 0.55 0.097 0.032 0) *
## 7089) reimbursement2008< 2535 733 423 B1 (0.42 0.38 0.14 0.057 0.0041)
## 14178) age>=97.5 10 3 B1 (0.7 0.2 0.1 0 0) *
## 14179) age< 97.5 723 420 B1 (0.42 0.38 0.14 0.058 0.0041)
## 28358) age< 90.5 689 394 B1 (0.43 0.38 0.13 0.055 0.0044)
## 56716) heart.failure< 0.5 367 198 B1 (0.46 0.36 0.14 0.035 0.0082) *
## 56717) heart.failure>=0.5 322 192 B2 (0.39 0.4 0.13 0.078 0)
## 113434) age< 67.5 78 43 B1 (0.45 0.29 0.19 0.064 0) *
## 113435) age>=67.5 244 137 B2 (0.37 0.44 0.11 0.082 0) *
## 28359) age>=90.5 34 18 B2 (0.24 0.47 0.18 0.12 0) *
## 3545) cancer>=0.5 49 29 B2 (0.33 0.41 0.24 0.02 0) *
## 1773) stroke>=0.5 44 20 B2 (0.25 0.55 0.11 0.068 0.023) *
## 887) kidney>=0.5 432 254 B2 (0.33 0.41 0.17 0.079 0.0093)
## 1774) reimbursement2008>=2215 403 232 B2 (0.32 0.42 0.17 0.069 0.0099) *
## 1775) reimbursement2008< 2215 29 16 B1 (0.45 0.24 0.1 0.21 0) *
## 111) reimbursement2008>=2665 1603 878 B2 (0.32 0.45 0.16 0.058 0.0087) *
## 7) reimbursement2008>=3065 69518 41677 B2 (0.27 0.4 0.2 0.11 0.017)
## 14) diabetes< 0.5 15717 8966 B1 (0.43 0.35 0.15 0.064 0.0071)
## 28) cancer< 0.5 13123 7034 B1 (0.46 0.34 0.13 0.058 0.0065)
## 56) arthritis< 0.5 9625 4692 B1 (0.51 0.31 0.12 0.054 0.0058)
## 112) ihd< 0.5 3135 1246 B1 (0.6 0.26 0.095 0.036 0.0032)
## 224) depression< 0.5 2292 821 B1 (0.64 0.24 0.08 0.034 0.0044) *
## 225) depression>=0.5 843 425 B1 (0.5 0.33 0.14 0.04 0)
## 450) age< 92.5 810 398 B1 (0.51 0.32 0.13 0.041 0)
## 900) reimbursement2008>=11525 117 40 B1 (0.66 0.21 0.068 0.06 0) *
## 901) reimbursement2008< 11525 693 358 B1 (0.48 0.33 0.14 0.038 0)
## 1802) reimbursement2008< 11105 684 352 B1 (0.49 0.34 0.14 0.038 0)
## 3604) reimbursement2008< 4365 286 134 B1 (0.53 0.33 0.12 0.017 0) *
## 3605) reimbursement2008>=4365 398 218 B1 (0.45 0.34 0.15 0.053 0)
## 7210) reimbursement2008>=4700 340 173 B1 (0.49 0.31 0.14 0.053 0) *
## 7211) reimbursement2008< 4700 58 28 B2 (0.22 0.52 0.21 0.052 0) *
## 1803) reimbursement2008>=11105 9 4 B3 (0.33 0.11 0.56 0 0) *
## 451) age>=92.5 33 14 B2 (0.18 0.58 0.21 0.03 0) *
## 113) ihd>=0.5 6490 3446 B1 (0.47 0.33 0.13 0.063 0.0071)
## 226) depression< 0.5 4266 2110 B1 (0.51 0.31 0.12 0.056 0.0061)
## 452) osteoporosis< 0.5 3304 1572 B1 (0.52 0.3 0.12 0.055 0.0064)
## 904) reimbursement2008>=5905 1626 714 B1 (0.56 0.25 0.12 0.061 0.0068) *
## 905) reimbursement2008< 5905 1678 858 B1 (0.49 0.34 0.12 0.05 0.006)
## 1810) reimbursement2008< 5695 1608 814 B1 (0.49 0.33 0.12 0.051 0.0062) *
## 1811) reimbursement2008>=5695 70 34 B2 (0.37 0.51 0.086 0.029 0) *
## 453) osteoporosis>=0.5 962 538 B1 (0.44 0.38 0.12 0.057 0.0052)
## 906) stroke< 0.5 857 465 B1 (0.46 0.37 0.12 0.056 0.0047)
## 1812) heart.failure< 0.5 405 203 B1 (0.5 0.35 0.1 0.044 0.0074)
## 3624) age< 83.5 329 159 B1 (0.52 0.32 0.11 0.049 0.0091) *
## 3625) age>=83.5 76 41 B2 (0.42 0.46 0.092 0.026 0)
## 7250) reimbursement2008>=6785 21 7 B1 (0.67 0.24 0.095 0 0) *
## 7251) reimbursement2008< 6785 55 25 B2 (0.33 0.55 0.091 0.036 0) *
## 1813) heart.failure>=0.5 452 262 B1 (0.42 0.38 0.13 0.066 0.0022)
## 3626) reimbursement2008>=3875 362 201 B1 (0.44 0.35 0.13 0.069 0.0028) *
## 3627) reimbursement2008< 3875 90 45 B2 (0.32 0.5 0.12 0.056 0)
## 7254) age< 69.5 21 9 B1 (0.57 0.29 0.048 0.095 0) *
## 7255) age>=69.5 69 30 B2 (0.25 0.57 0.14 0.043 0) *
## 907) stroke>=0.5 105 54 B2 (0.3 0.49 0.13 0.067 0.0095) *
## 227) depression>=0.5 2224 1336 B1 (0.4 0.35 0.16 0.076 0.009)
## 454) kidney< 0.5 1518 863 B1 (0.43 0.34 0.16 0.061 0.0053) *
## 455) kidney>=0.5 706 440 B2 (0.33 0.38 0.17 0.11 0.017)
## 910) reimbursement2008>=3155 696 431 B2 (0.33 0.38 0.16 0.11 0.017)
## 1820) heart.failure< 0.5 177 99 B1 (0.44 0.35 0.15 0.062 0) *
## 1821) heart.failure>=0.5 519 316 B2 (0.3 0.39 0.17 0.12 0.023) *
## 911) reimbursement2008< 3155 10 4 B3 (0.1 0.1 0.6 0.2 0) *
## 57) arthritis>=0.5 3498 2017 B2 (0.33 0.42 0.17 0.069 0.0083)
## 114) reimbursement2008< 8525 2340 1270 B2 (0.31 0.46 0.17 0.062 0.0064)
## 228) reimbursement2008< 4645 1359 754 B2 (0.34 0.45 0.15 0.056 0.0059)
## 456) ihd< 0.5 440 248 B2 (0.4 0.44 0.11 0.045 0.0045)
## 912) reimbursement2008< 3155 58 22 B2 (0.34 0.62 0 0.017 0.017) *
## 913) reimbursement2008>=3155 382 225 B1 (0.41 0.41 0.13 0.05 0.0026)
## 1826) reimbursement2008< 3245 25 8 B1 (0.68 0.28 0.04 0 0) *
## 1827) reimbursement2008>=3245 357 208 B2 (0.39 0.42 0.13 0.053 0.0028)
## 3654) age>=80.5 91 46 B1 (0.49 0.34 0.099 0.066 0) *
## 3655) age< 80.5 266 148 B2 (0.36 0.44 0.15 0.049 0.0038) *
## 457) ihd>=0.5 919 506 B2 (0.32 0.45 0.17 0.061 0.0065) *
## 229) reimbursement2008>=4645 981 516 B2 (0.26 0.47 0.19 0.069 0.0071) *
## 115) reimbursement2008>=8525 1158 722 B1 (0.38 0.35 0.17 0.085 0.012)
## 230) copd< 0.5 714 396 B1 (0.45 0.33 0.13 0.085 0.007)
## 460) depression< 0.5 412 196 B1 (0.52 0.29 0.1 0.073 0.0097) *
## 461) depression>=0.5 302 183 B2 (0.34 0.39 0.16 0.1 0.0033)
## 922) age>=92.5 9 3 B1 (0.67 0 0.11 0.11 0.11) *
## 923) age< 92.5 293 174 B2 (0.33 0.41 0.16 0.1 0)
## 1846) stroke>=0.5 39 19 B1 (0.51 0.31 0.1 0.077 0) *
## 1847) stroke< 0.5 254 147 B2 (0.3 0.42 0.17 0.11 0) *
## 231) copd>=0.5 444 272 B2 (0.27 0.39 0.24 0.086 0.02)
## 462) osteoporosis< 0.5 282 187 B2 (0.31 0.34 0.25 0.082 0.018)
## 924) reimbursement2008< 27390 220 143 B1 (0.35 0.3 0.26 0.073 0.018)
## 1848) reimbursement2008>=12810 132 78 B1 (0.41 0.32 0.2 0.068 0.0076)
## 3696) age< 84.5 105 55 B1 (0.48 0.33 0.15 0.029 0.0095) *
## 3697) age>=84.5 27 17 B3 (0.15 0.26 0.37 0.22 0) *
## 1849) reimbursement2008< 12810 88 57 B3 (0.26 0.27 0.35 0.08 0.034) *
## 925) reimbursement2008>=27390 62 33 B2 (0.18 0.47 0.23 0.11 0.016) *
## 463) osteoporosis>=0.5 162 85 B2 (0.19 0.48 0.22 0.093 0.025) *
## 29) cancer>=0.5 2594 1539 B2 (0.26 0.41 0.24 0.091 0.01)
## 58) reimbursement2008< 5770 1000 562 B2 (0.3 0.44 0.19 0.07 0.005) *
## 59) reimbursement2008>=5770 1594 977 B2 (0.23 0.39 0.27 0.1 0.014)
## 118) reimbursement2008>=8645 1054 656 B2 (0.27 0.38 0.24 0.1 0.015)
## 236) arthritis< 0.5 745 464 B2 (0.31 0.38 0.2 0.097 0.013)
## 472) ihd< 0.5 159 94 B1 (0.41 0.32 0.21 0.05 0.013)
## 944) reimbursement2008>=11995 76 36 B1 (0.53 0.24 0.16 0.066 0.013) *
## 945) reimbursement2008< 11995 83 50 B2 (0.3 0.4 0.25 0.036 0.012) *
## 473) ihd>=0.5 586 356 B2 (0.28 0.39 0.2 0.11 0.014) *
## 237) arthritis>=0.5 309 192 B2 (0.16 0.38 0.32 0.12 0.019)
## 474) reimbursement2008>=10960 237 136 B2 (0.15 0.43 0.3 0.11 0.013)
## 948) copd< 0.5 126 64 B2 (0.17 0.49 0.26 0.071 0) *
## 949) copd>=0.5 111 72 B2 (0.12 0.35 0.35 0.15 0.027)
## 1898) age< 75.5 54 30 B3 (0.15 0.26 0.44 0.13 0.019) *
## 1899) age>=75.5 57 32 B2 (0.088 0.44 0.26 0.18 0.035) *
## 475) reimbursement2008< 10960 72 44 B3 (0.19 0.22 0.39 0.15 0.042) *
## 119) reimbursement2008< 8645 540 321 B2 (0.16 0.41 0.32 0.11 0.011)
## 238) heart.failure>=0.5 243 128 B2 (0.14 0.47 0.28 0.099 0.016) *
## 239) heart.failure< 0.5 297 191 B3 (0.18 0.35 0.36 0.11 0.0067)
## 478) depression< 0.5 226 141 B2 (0.18 0.38 0.33 0.12 0.0044) *
## 479) depression>=0.5 71 39 B3 (0.17 0.27 0.45 0.099 0.014) *
## 15) diabetes>=0.5 53801 31450 B2 (0.23 0.42 0.21 0.13 0.02)
## 30) kidney< 0.5 25067 14311 B2 (0.3 0.43 0.19 0.076 0.0074)
## 60) arthritis< 0.5 15178 9179 B2 (0.35 0.4 0.17 0.069 0.0063)
## 120) cancer< 0.5 12572 7709 B2 (0.39 0.39 0.16 0.063 0.0059)
## 240) ihd< 0.5 2617 1376 B1 (0.47 0.34 0.13 0.049 0.0053)
## 480) reimbursement2008>=9400 403 171 B1 (0.58 0.21 0.15 0.05 0.0099) *
## 481) reimbursement2008< 9400 2214 1205 B1 (0.46 0.36 0.13 0.048 0.0045)
## 962) osteoporosis< 0.5 1636 847 B1 (0.48 0.34 0.12 0.049 0.0043)
## 1924) alzheimers< 0.5 1127 559 B1 (0.5 0.33 0.12 0.042 0.0035) *
## 1925) alzheimers>=0.5 509 288 B1 (0.43 0.36 0.13 0.065 0.0059)
## 3850) reimbursement2008< 3775 137 68 B1 (0.5 0.3 0.12 0.066 0.0073) *
## 3851) reimbursement2008>=3775 372 220 B1 (0.41 0.39 0.13 0.065 0.0054)
## 7702) reimbursement2008>=4055 330 188 B1 (0.43 0.36 0.13 0.07 0.0061)
## 15404) reimbursement2008>=4185 309 177 B1 (0.43 0.38 0.12 0.065 0.0065)
## 30808) reimbursement2008>=4635 253 138 B1 (0.45 0.35 0.12 0.067 0.0079)
## 61616) age< 96 245 131 B1 (0.47 0.36 0.11 0.065 0.0041)
## 123232) reimbursement2008< 8170 209 106 B1 (0.49 0.33 0.11 0.067 0) *
## 123233) reimbursement2008>=8170 36 19 B2 (0.31 0.47 0.14 0.056 0.028)
## 246466) age< 74.5 15 6 B1 (0.6 0.13 0.2 0 0.067) *
## 246467) age>=74.5 21 6 B2 (0.095 0.71 0.095 0.095 0) *
## 61617) age>=96 8 5 B3 (0.12 0.25 0.38 0.12 0.12) *
## 30809) reimbursement2008< 4635 56 28 B2 (0.3 0.5 0.14 0.054 0) *
## 15405) reimbursement2008< 4185 21 11 B1 (0.48 0.095 0.29 0.14 0) *
## 7703) reimbursement2008< 4055 42 17 B2 (0.24 0.6 0.14 0.024 0) *
## 963) osteoporosis>=0.5 578 342 B2 (0.38 0.41 0.16 0.047 0.0052)
## 1926) depression< 0.5 339 189 B1 (0.44 0.37 0.13 0.047 0.0029)
## 3852) reimbursement2008< 4905 211 119 B1 (0.44 0.42 0.11 0.033 0)
## 7704) reimbursement2008< 4075 142 71 B1 (0.5 0.36 0.11 0.035 0) *
## 7705) reimbursement2008>=4075 69 31 B2 (0.3 0.55 0.12 0.029 0) *
## 3853) reimbursement2008>=4905 128 70 B1 (0.45 0.3 0.17 0.07 0.0078) *
## 1927) depression>=0.5 239 130 B2 (0.29 0.46 0.2 0.046 0.0084)
## 3854) copd< 0.5 181 88 B2 (0.31 0.51 0.13 0.039 0.011) *
## 3855) copd>=0.5 58 34 B3 (0.24 0.28 0.41 0.069 0) *
## 241) ihd>=0.5 9955 5976 B2 (0.36 0.4 0.17 0.067 0.006)
## 482) depression< 0.5 5563 3339 B1 (0.4 0.38 0.15 0.059 0.0059)
## 964) reimbursement2008>=8955 1363 758 B1 (0.44 0.32 0.16 0.067 0.0088)
## 1928) copd< 0.5 798 405 B1 (0.49 0.32 0.12 0.064 0.0038) *
## 1929) copd>=0.5 565 353 B1 (0.38 0.32 0.22 0.073 0.016)
## 3858) stroke>=0.5 116 64 B2 (0.35 0.45 0.12 0.06 0.017)
## 7716) age>=74.5 63 36 B1 (0.43 0.33 0.16 0.063 0.016) *
## 7717) age< 74.5 53 22 B2 (0.26 0.58 0.075 0.057 0.019) *
## 3859) stroke< 0.5 449 278 B1 (0.38 0.28 0.24 0.076 0.016) *
## 965) reimbursement2008< 8955 4200 2510 B2 (0.39 0.4 0.15 0.056 0.005)
## 1930) heart.failure< 0.5 1953 1129 B1 (0.42 0.4 0.13 0.045 0.0041)
## 3860) reimbursement2008< 3415 343 172 B1 (0.5 0.37 0.096 0.032 0.0058) *
## 3861) reimbursement2008>=3415 1610 954 B2 (0.41 0.41 0.14 0.048 0.0037)
## 7722) age< 42.5 43 17 B1 (0.6 0.26 0.07 0.07 0) *
## 7723) age>=42.5 1567 922 B2 (0.4 0.41 0.14 0.047 0.0038)
## 15446) age>=50.5 1527 894 B2 (0.4 0.41 0.13 0.047 0.0039)
## 30892) reimbursement2008>=3465 1478 870 B2 (0.41 0.41 0.13 0.045 0.0027)
## 61784) reimbursement2008< 4655 759 431 B1 (0.43 0.4 0.12 0.043 0.0013)
## 123568) reimbursement2008>=4315 158 79 B1 (0.5 0.35 0.095 0.051 0.0063) *
## 123569) reimbursement2008< 4315 601 352 B1 (0.41 0.41 0.13 0.042 0)
## 247138) reimbursement2008< 4295 592 344 B1 (0.42 0.41 0.13 0.042 0)
## 494276) age>=82.5 135 71 B1 (0.47 0.36 0.15 0.015 0) *
## 494277) age< 82.5 457 266 B2 (0.4 0.42 0.13 0.05 0)
## 988554) age< 74.5 290 166 B1 (0.43 0.39 0.13 0.052 0)
## 1977108) age>=62.5 234 128 B1 (0.45 0.38 0.12 0.047 0) *
## 1977109) age< 62.5 56 31 B2 (0.32 0.45 0.16 0.071 0) *
## 988555) age>=74.5 167 90 B2 (0.36 0.46 0.13 0.048 0)
## 1977110) reimbursement2008>=4105 39 20 B1 (0.49 0.36 0.13 0.026 0) *
## 1977111) reimbursement2008< 4105 128 65 B2 (0.32 0.49 0.13 0.055 0) *
## 247139) reimbursement2008>=4295 9 1 B2 (0.11 0.89 0 0 0) *
## 61785) reimbursement2008>=4655 719 414 B2 (0.38 0.42 0.15 0.047 0.0042)
## 123570) reimbursement2008< 5835 346 180 B2 (0.35 0.48 0.13 0.038 0.0029) *
## 123571) reimbursement2008>=5835 373 223 B1 (0.4 0.37 0.16 0.056 0.0054)
## 247142) alzheimers>=0.5 124 64 B1 (0.48 0.31 0.15 0.04 0.0081)
## 494284) reimbursement2008< 8555 114 55 B1 (0.52 0.28 0.16 0.035 0.0088) *
## 494285) reimbursement2008>=8555 10 3 B2 (0.1 0.7 0.1 0.1 0) *
## 247143) alzheimers< 0.5 249 149 B2 (0.36 0.4 0.17 0.064 0.004)
## 494286) reimbursement2008>=6045 217 124 B2 (0.36 0.43 0.14 0.069 0.0046) *
## 494287) reimbursement2008< 6045 32 20 B1 (0.38 0.22 0.38 0.031 0)
## 988574) age< 72.5 11 4 B1 (0.64 0.18 0.18 0 0) *
## 988575) age>=72.5 21 11 B3 (0.24 0.24 0.48 0.048 0) *
## 30893) reimbursement2008< 3465 49 24 B2 (0.27 0.51 0.082 0.1 0.041) *
## 15447) age< 50.5 40 26 B1 (0.35 0.3 0.3 0.05 0) *
## 1931) heart.failure>=0.5 2247 1339 B2 (0.35 0.4 0.17 0.066 0.0058)
## 3862) reimbursement2008>=5335 866 530 B1 (0.39 0.37 0.16 0.074 0.0058)
## 7724) reimbursement2008>=8115 129 68 B2 (0.36 0.47 0.12 0.047 0) *
## 7725) reimbursement2008< 8115 737 447 B1 (0.39 0.35 0.17 0.079 0.0068)
## 15450) age< 94.5 703 421 B1 (0.4 0.35 0.17 0.075 0.0071)
## 30900) reimbursement2008>=6635 298 164 B1 (0.45 0.32 0.15 0.067 0.013) *
## 30901) reimbursement2008< 6635 405 255 B2 (0.37 0.37 0.18 0.081 0.0025)
## 61802) reimbursement2008< 5685 137 80 B1 (0.42 0.31 0.16 0.11 0) *
## 61803) reimbursement2008>=5685 268 161 B2 (0.34 0.4 0.19 0.067 0.0037) *
## 15451) age>=94.5 34 17 B2 (0.24 0.5 0.12 0.15 0) *
## 3863) reimbursement2008< 5335 1381 795 B2 (0.33 0.42 0.18 0.061 0.0058)
## 7726) copd< 0.5 997 591 B2 (0.36 0.41 0.17 0.057 0.006)
## 15452) age< 69.5 297 171 B1 (0.42 0.38 0.15 0.04 0.0034)
## 30904) reimbursement2008< 5065 274 153 B1 (0.44 0.36 0.15 0.04 0.0036)
## 61808) alzheimers< 0.5 174 91 B1 (0.48 0.3 0.17 0.046 0.0057) *
## 61809) alzheimers>=0.5 100 54 B2 (0.38 0.46 0.13 0.03 0)
## 123618) reimbursement2008>=4355 26 10 B1 (0.62 0.15 0.15 0.077 0) *
## 123619) reimbursement2008< 4355 74 32 B2 (0.3 0.57 0.12 0.014 0) *
## 30905) reimbursement2008>=5065 23 9 B2 (0.22 0.61 0.13 0.043 0) *
## 15453) age>=69.5 700 407 B2 (0.33 0.42 0.18 0.064 0.0071) *
## 7727) copd>=0.5 384 204 B2 (0.27 0.47 0.19 0.07 0.0052) *
## 483) depression>=0.5 4392 2538 B2 (0.31 0.42 0.18 0.077 0.0061)
## 966) reimbursement2008< 8325 2928 1619 B2 (0.31 0.45 0.17 0.065 0.0065)
## 1932) copd< 0.5 1987 1102 B2 (0.33 0.45 0.16 0.056 0.0045)
## 3864) age< 98.5 1964 1085 B2 (0.33 0.45 0.16 0.057 0.0046)
## 7728) reimbursement2008< 3085 22 8 B1 (0.64 0.23 0.045 0.091 0) *
## 7729) reimbursement2008>=3085 1942 1068 B2 (0.33 0.45 0.16 0.056 0.0046)
## 15458) heart.failure< 0.5 889 508 B2 (0.37 0.43 0.15 0.051 0.0034) *
## 15459) heart.failure>=0.5 1053 560 B2 (0.3 0.47 0.17 0.061 0.0057)
## 30918) osteoporosis< 0.5 721 396 B2 (0.32 0.45 0.16 0.061 0.0055)
## 61836) age>=86.5 109 59 B1 (0.46 0.3 0.15 0.083 0.0092) *
## 61837) age< 86.5 612 320 B2 (0.3 0.48 0.16 0.057 0.0049) *
## 30919) osteoporosis>=0.5 332 164 B2 (0.24 0.51 0.18 0.06 0.006) *
## 3865) age>=98.5 23 12 B3 (0.26 0.26 0.48 0 0) *
## 1933) copd>=0.5 941 517 B2 (0.26 0.45 0.2 0.084 0.011) *
## 967) reimbursement2008>=8325 1464 919 B2 (0.32 0.37 0.2 0.1 0.0055)
## 1934) reimbursement2008< 8485 36 16 B1 (0.56 0.22 0.22 0 0) *
## 1935) reimbursement2008>=8485 1428 891 B2 (0.32 0.38 0.2 0.1 0.0056)
## 3870) age< 78.5 837 532 B2 (0.35 0.36 0.19 0.098 0.0036)
## 7740) reimbursement2008< 21320 639 406 B1 (0.36 0.35 0.19 0.092 0.0031)
## 15480) age< 49.5 83 47 B2 (0.35 0.43 0.096 0.12 0) *
## 15481) age>=49.5 556 352 B1 (0.37 0.33 0.21 0.088 0.0036)
## 30962) age>=67.5 368 230 B1 (0.38 0.36 0.18 0.087 0)
## 61924) reimbursement2008>=10440 261 153 B1 (0.41 0.33 0.17 0.084 0)
## 123848) reimbursement2008< 12585 92 46 B1 (0.5 0.25 0.18 0.065 0) *
## 123849) reimbursement2008>=12585 169 105 B2 (0.37 0.38 0.16 0.095 0)
## 247698) reimbursement2008>=14485 109 63 B1 (0.42 0.31 0.17 0.092 0)
## 495396) osteoporosis< 0.5 72 37 B1 (0.49 0.24 0.17 0.11 0) *
## 495397) osteoporosis>=0.5 37 20 B2 (0.3 0.46 0.19 0.054 0) *
## 247699) reimbursement2008< 14485 60 30 B2 (0.27 0.5 0.13 0.1 0) *
## 61925) reimbursement2008< 10440 107 62 B2 (0.28 0.42 0.21 0.093 0) *
## 30963) age< 67.5 188 122 B1 (0.35 0.28 0.27 0.09 0.011)
## 61926) age>=55.5 135 82 B1 (0.39 0.25 0.27 0.089 0) *
## 61927) age< 55.5 53 34 B2 (0.25 0.36 0.26 0.094 0.038) *
## 7741) reimbursement2008>=21320 198 114 B2 (0.3 0.42 0.16 0.12 0.0051) *
## 3871) age>=78.5 591 359 B2 (0.28 0.39 0.21 0.11 0.0085)
## 7742) heart.failure< 0.5 122 73 B1 (0.4 0.32 0.18 0.098 0)
## 15484) reimbursement2008< 11560 40 17 B1 (0.58 0.2 0.18 0.05 0) *
## 15485) reimbursement2008>=11560 82 51 B2 (0.32 0.38 0.18 0.12 0) *
## 7743) heart.failure>=0.5 469 276 B2 (0.24 0.41 0.22 0.11 0.011) *
## 121) cancer>=0.5 2606 1470 B2 (0.21 0.44 0.25 0.098 0.0081) *
## 61) arthritis>=0.5 9889 5132 B2 (0.22 0.48 0.21 0.088 0.0092)
## 122) depression< 0.5 5134 2665 B2 (0.25 0.48 0.18 0.08 0.0078)
## 244) cancer< 0.5 4305 2260 B2 (0.27 0.48 0.17 0.076 0.0086)
## 488) reimbursement2008>=9880 1063 636 B2 (0.32 0.4 0.18 0.089 0.012)
## 976) ihd< 0.5 102 49 B1 (0.52 0.27 0.13 0.069 0.0098) *
## 977) ihd>=0.5 961 562 B2 (0.29 0.42 0.19 0.092 0.012) *
## 489) reimbursement2008< 9880 3242 1624 B2 (0.25 0.5 0.17 0.072 0.0074) *
## 245) cancer>=0.5 829 405 B2 (0.15 0.51 0.23 0.1 0.0036) *
## 123) depression>=0.5 4755 2467 B2 (0.18 0.48 0.23 0.096 0.011) *
## 31) kidney>=0.5 28734 17139 B2 (0.16 0.4 0.23 0.17 0.03)
## 62) reimbursement2008< 15395 16249 9131 B2 (0.19 0.44 0.24 0.12 0.016)
## 124) arthritis< 0.5 9424 5647 B2 (0.23 0.4 0.23 0.12 0.017)
## 248) cancer< 0.5 7786 4711 B2 (0.25 0.39 0.21 0.12 0.017)
## 496) ihd< 0.5 964 608 B1 (0.37 0.36 0.17 0.085 0.011)
## 992) depression< 0.5 572 338 B1 (0.41 0.32 0.16 0.1 0.01)
## 1984) reimbursement2008< 3545 101 53 B2 (0.33 0.48 0.15 0.04 0.0099) *
## 1985) reimbursement2008>=3545 471 270 B1 (0.43 0.29 0.16 0.11 0.011)
## 3970) osteoporosis< 0.5 346 186 B1 (0.46 0.27 0.15 0.11 0.014) *
## 3971) osteoporosis>=0.5 125 82 B2 (0.33 0.34 0.2 0.13 0)
## 7942) age>=62 106 67 B1 (0.37 0.37 0.15 0.11 0)
## 15884) age>=67.5 93 55 B2 (0.34 0.41 0.16 0.086 0)
## 31768) reimbursement2008>=6110 44 23 B1 (0.48 0.3 0.11 0.11 0)
## 63536) reimbursement2008< 9180 26 9 B1 (0.65 0.15 0.077 0.12 0) *
## 63537) reimbursement2008>=9180 18 9 B2 (0.22 0.5 0.17 0.11 0) *
## 31769) reimbursement2008< 6110 49 24 B2 (0.22 0.51 0.2 0.061 0) *
## 15885) age< 67.5 13 6 B1 (0.54 0.077 0.077 0.31 0) *
## 7943) age< 62 19 10 B3 (0.11 0.21 0.47 0.21 0) *
## 993) depression>=0.5 392 227 B2 (0.31 0.42 0.19 0.064 0.013)
## 1986) reimbursement2008>=14460 9 2 B1 (0.78 0.22 0 0 0) *
## 1987) reimbursement2008< 14460 383 220 B2 (0.3 0.43 0.2 0.065 0.013) *
## 497) ihd>=0.5 6822 4095 B2 (0.24 0.4 0.22 0.12 0.018)
## 994) reimbursement2008< 6325 3172 1786 B2 (0.22 0.44 0.22 0.11 0.016) *
## 995) reimbursement2008>=6325 3650 2309 B2 (0.25 0.37 0.22 0.14 0.019)
## 1990) osteoporosis< 0.5 2424 1594 B2 (0.27 0.34 0.23 0.14 0.02)
## 3980) depression< 0.5 1234 816 B2 (0.3 0.34 0.2 0.13 0.024)
## 7960) reimbursement2008>=12135 349 226 B1 (0.35 0.3 0.16 0.16 0.032)
## 15920) age>=54 331 210 B1 (0.37 0.28 0.16 0.16 0.03) *
## 15921) age< 54 18 9 B2 (0.11 0.5 0.11 0.22 0.056) *
## 7961) reimbursement2008< 12135 885 570 B2 (0.28 0.36 0.22 0.12 0.021) *
## 3981) depression>=0.5 1190 778 B2 (0.24 0.35 0.25 0.15 0.016)
## 7962) copd< 0.5 547 367 B2 (0.28 0.33 0.25 0.12 0.022)
## 15924) reimbursement2008>=9205 310 209 B1 (0.33 0.32 0.21 0.12 0.029)
## 31848) reimbursement2008< 9955 50 28 B2 (0.42 0.44 0.02 0.12 0) *
## 31849) reimbursement2008>=9955 260 180 B1 (0.31 0.3 0.24 0.12 0.035)
## 63698) reimbursement2008>=14765 20 9 B1 (0.55 0.25 0.1 0.05 0.05) *
## 63699) reimbursement2008< 14765 240 168 B2 (0.29 0.3 0.25 0.12 0.033)
## 127398) age>=61.5 201 138 B1 (0.31 0.29 0.24 0.13 0.02)
## 254796) reimbursement2008< 12625 112 69 B1 (0.38 0.24 0.28 0.089 0.0089) *
## 254797) reimbursement2008>=12625 89 58 B2 (0.22 0.35 0.2 0.19 0.034) *
## 127399) age< 61.5 39 25 B2 (0.15 0.36 0.31 0.077 0.1) *
## 15925) reimbursement2008< 9205 237 156 B2 (0.22 0.34 0.3 0.12 0.013)
## 31850) age< 67.5 56 29 B2 (0.2 0.48 0.16 0.16 0) *
## 31851) age>=67.5 181 118 B3 (0.23 0.3 0.35 0.1 0.017)
## 63702) reimbursement2008>=6865 136 82 B3 (0.25 0.26 0.4 0.074 0.015) *
## 63703) reimbursement2008< 6865 45 27 B2 (0.18 0.4 0.2 0.2 0.022) *
## 7963) copd>=0.5 643 411 B2 (0.21 0.36 0.25 0.17 0.011) *
## 1991) osteoporosis>=0.5 1226 715 B2 (0.21 0.42 0.22 0.14 0.017) *
## 249) cancer>=0.5 1638 936 B2 (0.13 0.43 0.29 0.14 0.016) *
## 125) arthritis>=0.5 6825 3484 B2 (0.13 0.49 0.25 0.12 0.014) *
## 63) reimbursement2008>=15395 12485 8008 B2 (0.13 0.36 0.23 0.24 0.049)
## 126) arthritis>=0.5 5402 3220 B2 (0.094 0.4 0.24 0.22 0.04)
## 252) reimbursement2008< 34925 3345 1942 B2 (0.11 0.42 0.25 0.19 0.03)
## 504) depression< 0.5 1291 714 B2 (0.14 0.45 0.22 0.17 0.025)
## 1008) cancer< 0.5 973 546 B2 (0.16 0.44 0.19 0.18 0.029) *
## 1009) cancer>=0.5 318 168 B2 (0.072 0.47 0.3 0.14 0.013)
## 2018) reimbursement2008>=16525 293 144 B2 (0.068 0.51 0.28 0.13 0.014) *
## 2019) reimbursement2008< 16525 25 10 B3 (0.12 0.04 0.6 0.24 0) *
## 505) depression>=0.5 2054 1228 B2 (0.092 0.4 0.27 0.2 0.034) *
## 253) reimbursement2008>=34925 2057 1278 B2 (0.067 0.38 0.23 0.27 0.055)
## 506) copd< 0.5 520 300 B2 (0.096 0.42 0.23 0.21 0.042) *
## 507) copd>=0.5 1537 978 B2 (0.057 0.36 0.23 0.29 0.06)
## 1014) age>=62.5 1286 804 B2 (0.058 0.37 0.24 0.27 0.061) *
## 1015) age< 62.5 251 153 B4 (0.052 0.31 0.2 0.39 0.052)
## 2030) reimbursement2008< 101585 237 150 B4 (0.055 0.32 0.21 0.37 0.051)
## 4060) cancer>=0.5 62 36 B2 (0.048 0.42 0.27 0.24 0.016) *
## 4061) cancer< 0.5 175 103 B4 (0.057 0.29 0.18 0.41 0.063) *
## 2031) reimbursement2008>=101585 14 3 B4 (0 0.071 0.071 0.79 0.071) *
## 127) arthritis< 0.5 7083 4788 B2 (0.15 0.32 0.22 0.25 0.057)
## 254) cancer< 0.5 5298 3651 B2 (0.17 0.31 0.2 0.26 0.062)
## 508) depression< 0.5 2489 1797 B2 (0.22 0.28 0.18 0.27 0.06)
## 1016) copd>=0.5 1317 890 B2 (0.2 0.32 0.18 0.24 0.056)
## 2032) ihd< 0.5 72 41 B1 (0.43 0.25 0.15 0.11 0.056) *
## 2033) ihd>=0.5 1245 836 B2 (0.19 0.33 0.18 0.24 0.056) *
## 1017) copd< 0.5 1172 815 B4 (0.23 0.23 0.17 0.3 0.065)
## 2034) reimbursement2008>=43640 191 129 B2 (0.15 0.32 0.23 0.22 0.073)
## 4068) age>=64.5 172 112 B2 (0.16 0.35 0.2 0.23 0.064) *
## 4069) age< 64.5 19 10 B3 (0.11 0.11 0.47 0.16 0.16) *
## 2035) reimbursement2008< 43640 981 666 B4 (0.25 0.21 0.16 0.32 0.063)
## 4070) reimbursement2008< 23175 468 337 B1 (0.28 0.24 0.16 0.26 0.056)
## 8140) age< 93.5 457 326 B1 (0.29 0.23 0.16 0.26 0.057)
## 16280) age< 86.5 398 288 B1 (0.28 0.25 0.16 0.25 0.063)
## 32560) alzheimers>=0.5 179 122 B1 (0.32 0.28 0.15 0.17 0.073)
## 65120) reimbursement2008>=21440 38 19 B2 (0.24 0.5 0.11 0.13 0.026) *
## 65121) reimbursement2008< 21440 141 93 B1 (0.34 0.23 0.16 0.18 0.085)
## 130242) reimbursement2008>=17585 89 53 B1 (0.4 0.17 0.19 0.16 0.079) *
## 130243) reimbursement2008< 17585 52 35 B2 (0.23 0.33 0.12 0.23 0.096) *
## 32561) alzheimers< 0.5 219 151 B4 (0.24 0.23 0.16 0.31 0.055) *
## 16281) age>=86.5 59 38 B1 (0.36 0.1 0.17 0.36 0.017)
## 32562) reimbursement2008>=19680 18 8 B1 (0.56 0.11 0 0.28 0.056) *
## 32563) reimbursement2008< 19680 41 25 B4 (0.27 0.098 0.24 0.39 0) *
## 8141) age>=93.5 11 6 B2 (0 0.45 0.36 0.18 0) *
## 4071) reimbursement2008>=23175 513 320 B4 (0.22 0.18 0.16 0.38 0.07) *
## 509) depression>=0.5 2809 1854 B2 (0.13 0.34 0.22 0.25 0.063) *
## 255) cancer>=0.5 1785 1137 B2 (0.097 0.36 0.28 0.22 0.041) *## [1] TRUEreplay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"model.selected")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
## 2.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction data.new.prediction firing: model.selected
## 3.0000 3 0 2 1 0
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="fit.data.training.all",
chunk_step_major=max(glb_script_df$chunk_step_major)+1,
chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed8 fit.models 5 0 71.848
## elapsed9 fit.data.training.all 6 0 9768.275Step 6: fit.data.training.all
if (glb_fin_mdl_id %in% names(glb_models_lst)) {
warning("Final model same as user selected model")
glb_fin_mdl <- glb_sel_mdl
} else {
print(mdl_feats_df <- myextract_mdl_feats( sel_mdl=glb_sel_mdl,
entity_df=glb_entity_df))
# Sync with parameters in mydsutils.R
ret_lst <- myfit_mdl_fn(model_id="Final",
indep_vars_vctr=mdl_feats_df$id,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_entity_df,
model_method=glb_sel_mdl$method,
model_loss_mtrx=glb_model_metric_terms, # Automate this
model_summaryFunction=glb_sel_mdl$control$summaryFunction,
model_metric=glb_sel_mdl$metric,
model_metric_maximize=glb_sel_mdl$maximize)
glb_fin_mdl <- glb_models_lst[["Final"]]
}## Warning: Final model same as user selected modelglb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="fit.data.training.all",
chunk_step_major=glb_script_df[nrow(glb_script_df), "chunk_step_major"],
chunk_step_minor=glb_script_df[nrow(glb_script_df), "chunk_step_minor"]+1,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor elapsed
## elapsed9 fit.data.training.all 6 0 9768.275
## elapsed10 fit.data.training.all 6 1 9789.672if (glb_is_regression) {
glb_entity_df[, glb_rsp_var_out] <- predict(glb_fin_mdl, newdata=glb_entity_df)
print(myplot_scatter(glb_entity_df, glb_rsp_var, glb_rsp_var_out,
smooth=TRUE))
glb_entity_df[, paste0(glb_rsp_var_out, ".err")] <-
abs(glb_entity_df[, glb_rsp_var_out] - glb_entity_df[, glb_rsp_var])
print(head(orderBy(reformulate(c("-", paste0(glb_rsp_var_out, ".err"))),
glb_entity_df)))
}
if (glb_is_classification & glb_is_binomial) {
stop("not implemented")
if (any(class(glb_fin_mdl) %in% c("train"))) {
glb_entity_df[, paste0(glb_rsp_var_out, ".proba")] <-
predict(glb_fin_mdl, newdata=glb_entity_df, type="prob")[, 2]
} else if (any(class(glb_fin_mdl) %in% c("rpart", "randomForest"))) {
glb_entity_df[, paste0(glb_rsp_var_out, ".proba")] <-
predict(glb_fin_mdl, newdata=glb_entity_df, type="prob")[, 2]
} else if (class(glb_fin_mdl) == "glm") {
stop("not implemented yet")
glb_entity_df[, paste0(glb_rsp_var_out, ".proba")] <-
predict(glb_fin_mdl, newdata=glb_entity_df, type="response")
} else stop("not implemented yet")
require(ROCR)
ROCRpred <- prediction(glb_entity_df[, paste0(glb_rsp_var_out, ".proba")],
glb_entity_df[, glb_rsp_var])
ROCRperf <- performance(ROCRpred, "tpr", "fpr")
plot(ROCRperf, colorize=TRUE, print.cutoffs.at=seq(0, 1, 0.1), text.adj=c(-0.2,1.7))
thresholds_df <- data.frame(threshold=seq(0.0, 1.0, 0.1))
thresholds_df$f.score <- sapply(1:nrow(thresholds_df), function(row_ix)
mycompute_classifier_f.score(mdl=glb_fin_mdl, obs_df=glb_entity_df,
proba_threshold=thresholds_df[row_ix, "threshold"],
rsp_var=glb_rsp_var,
rsp_var_out=glb_rsp_var_out))
print(thresholds_df)
print(myplot_line(thresholds_df, "threshold", "f.score"))
proba_threshold <- thresholds_df[which.max(thresholds_df$f.score),
"threshold"]
# This should change to maximize f.score.OOB ???
print(sprintf("Classifier Probability Threshold: %0.4f to maximize f.score.fit",
proba_threshold))
if (is.null(glb_clf_proba_threshold))
glb_clf_proba_threshold <- proba_threshold else {
print(sprintf("Classifier Probability Threshold: %0.4f per user specs",
glb_clf_proba_threshold))
}
if ((class(glb_entity_df[, glb_rsp_var]) != "factor") |
(length(levels(glb_entity_df[, glb_rsp_var])) != 2))
stop("expecting a factor with two levels:", glb_rsp_var)
glb_entity_df[, glb_rsp_var_out] <-
factor(levels(glb_entity_df[, glb_rsp_var])[
(glb_entity_df[, paste0(glb_rsp_var_out, ".proba")] >=
glb_clf_proba_threshold) * 1 + 1])
print(mycreate_xtab(glb_entity_df, c(glb_rsp_var, glb_rsp_var_out)))
print(sprintf("f.score=%0.4f",
mycompute_classifier_f.score(glb_fin_mdl, glb_entity_df,
glb_clf_proba_threshold,
glb_rsp_var, glb_rsp_var_out)))
}
if (glb_is_classification & !glb_is_binomial) {
glb_entity_df[, glb_rsp_var_out] <- predict(glb_fin_mdl, newdata=glb_entity_df, type="raw")
}
print(glb_feats_df <- mymerge_feats_importance(feats_df=glb_feats_df, sel_mdl=glb_fin_mdl,
entity_df=glb_entity_df))## id cor.y exclude.as.feat cor.y.abs cor.low
## 16 reimbursement2008 0.3756473063 0 0.3756473063 0
## 5 bucket2008 0.4518935763 0 0.4518935763 1
## 11 diabetes 0.3904719536 0 0.3904719536 1
## 13 ihd 0.3905905884 0 0.3905905884 1
## 12 heart.failure 0.3647689526 0 0.3647689526 1
## 14 kidney 0.3683780944 0 0.3683780944 1
## 4 arthritis 0.2717113526 0 0.2717113526 1
## 10 depression 0.2835366153 0 0.2835366153 1
## 8 cancer 0.2100892954 0 0.2100892954 1
## 2 age 0.0495694151 0 0.0495694151 1
## 9 copd 0.3108325355 0 0.3108325355 1
## 15 osteoporosis 0.2076745377 0 0.2076745377 1
## 3 alzheimers 0.2741643394 0 0.2741643394 1
## 18 stroke 0.1846626746 0 0.1846626746 1
## 1 .rnorm -0.0007970401 0 0.0007970401 1
## 6 bucket2008.fctr 0.4518935763 1 0.4518935763 0
## 7 bucket2009 1.0000000000 1 1.0000000000 0
## 17 reimbursement2009 0.8581631864 1 0.8581631864 0
## importance
## 16 100.00000000
## 5 83.20987115
## 11 64.93876310
## 13 62.97689485
## 12 52.37678436
## 14 9.66286974
## 4 5.84119903
## 10 2.26938048
## 8 1.78931779
## 2 1.02716655
## 9 0.89198875
## 15 0.66926118
## 3 0.09536685
## 18 0.00000000
## 1 NA
## 6 NA
## 7 NA
## 17 NA# Most of this code is used again in predict.data.new chunk
glb_analytics_diag_plots <- function(obs_df) {
for (var in subset(glb_feats_df, !is.na(importance))$id) {
plot_df <- melt(obs_df, id.vars=var,
measure.vars=c(glb_rsp_var, glb_rsp_var_out))
# if (var == "<feat_name>") print(myplot_scatter(plot_df, var, "value",
# facet_colcol_name="variable") +
# geom_vline(xintercept=<divider_val>, linetype="dotted")) else
print(myplot_scatter(plot_df, var, "value", colorcol_name="variable",
facet_colcol_name="variable", jitter=TRUE) +
guides(color=FALSE))
}
if (glb_is_regression) {
plot_vars_df <- subset(glb_feats_df, Pr.z < 0.1)
print(myplot_prediction_regression(obs_df,
ifelse(nrow(plot_vars_df) > 1, plot_vars_df$id[2], ".rownames"),
plot_vars_df$id[1],
glb_rsp_var, glb_rsp_var_out)
# + facet_wrap(reformulate(plot_vars_df$id[2])) # if [1,2] is a factor
# + geom_point(aes_string(color="<col_name>.fctr")) # to color the plot
)
}
if (glb_is_classification) {
if (nrow(plot_vars_df <- subset(glb_feats_df, !is.na(importance))) == 0)
warning("No features in selected model are statistically important")
else print(myplot_prediction_classification(df=obs_df,
feat_x=ifelse(nrow(plot_vars_df) > 1, plot_vars_df$id[2],
".rownames"),
feat_y=plot_vars_df$id[1],
rsp_var=glb_rsp_var,
rsp_var_out=glb_rsp_var_out,
id_vars=glb_id_vars)
# + geom_hline(yintercept=<divider_val>, linetype = "dotted")
)
}
}
glb_analytics_diag_plots(obs_df=glb_entity_df)
## age alzheimers arthritis cancer copd depression diabetes
## 1 85 0 0 0 0 0 0
## 90001 78 0 0 1 0 1 1
## 90003 83 1 0 0 0 1 1
## 90005 91 0 0 0 0 1 1
## 90010 84 0 0 0 0 0 1
## 90491 73 0 0 0 1 0 0
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 1 0 0 0 0 0 0
## 90001 1 1 1 1 0 2720
## 90003 1 1 1 0 0 2740
## 90005 0 1 0 1 0 2810
## 90010 0 1 0 1 0 2970
## 90491 1 0 1 0 0 55000
## bucket2008 reimbursement2009 bucket2009 .rnorm bucket2009.fctr
## 1 1 0 1 -0.92480010 B1
## 90001 1 0 1 0.46373167 B1
## 90003 1 0 1 -0.19796903 B1
## 90005 1 0 1 0.92002065 B1
## 90010 1 0 1 0.04393011 B1
## 90491 5 0 1 1.17487693 B1
## bucket2008.fctr bucket2009.fctr.predict.
## 1 B1 B1
## 90001 B1 B2
## 90003 B1 B2
## 90005 B1 B2
## 90010 B1 B2
## 90491 B5 B1
## bucket2009.fctr.predict..accurate .label
## 1 TRUE .1
## 90001 FALSE .90001
## 90003 FALSE .90003
## 90005 FALSE .90005
## 90010 FALSE .90010
## 90491 TRUE .90491
## age alzheimers arthritis cancer copd depression diabetes
## 90003 83 1 0 0 0 1 1
## 105382 62 0 0 1 1 1 0
## 127997 62 1 0 1 1 1 1
## 307449 66 0 0 0 0 0 0
## 307451 86 0 0 0 0 0 0
## 366459 71 1 0 0 1 1 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 90003 1 1 1 0 0 2740
## 105382 1 1 1 0 0 65580
## 127997 1 1 0 0 1 60020
## 307449 0 0 0 0 0 0
## 307451 0 0 0 0 0 0
## 366459 1 1 1 0 0 194910
## bucket2008 reimbursement2009 bucket2009 .rnorm bucket2009.fctr
## 90003 1 0 1 -0.19796903 B1
## 105382 5 70 1 1.40881604 B1
## 127997 5 240 1 0.25106429 B1
## 307449 1 3000 2 0.03560744 B2
## 307451 1 3000 2 -1.58625862 B2
## 366459 5 5400 2 1.64137090 B2
## bucket2008.fctr bucket2009.fctr.predict.
## 90003 B1 B2
## 105382 B5 B2
## 127997 B5 B2
## 307449 B1 B1
## 307451 B1 B1
## 366459 B5 B2
## bucket2009.fctr.predict..accurate .label
## 90003 FALSE .90003
## 105382 FALSE .105382
## 127997 FALSE .127997
## 307449 FALSE .307449
## 307451 FALSE .307451
## 366459 TRUE .366459
## age alzheimers arthritis cancer copd depression diabetes
## 307446 66 0 0 0 0 0 0
## 307449 66 0 0 0 0 0 0
## 307450 76 0 0 0 0 0 0
## 307451 86 0 0 0 0 0 0
## 307452 68 0 0 0 0 0 0
## 366459 71 1 0 0 1 1 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 307446 0 0 0 0 0 0
## 307449 0 0 0 0 0 0
## 307450 0 0 0 0 0 0
## 307451 0 0 0 0 0 0
## 307452 0 0 0 0 0 0
## 366459 1 1 1 0 0 194910
## bucket2008 reimbursement2009 bucket2009 .rnorm bucket2009.fctr
## 307446 1 3000 2 0.32846944 B2
## 307449 1 3000 2 0.03560744 B2
## 307450 1 3000 2 -2.09348196 B2
## 307451 1 3000 2 -1.58625862 B2
## 307452 1 3000 2 -0.22660343 B2
## 366459 5 5400 2 1.64137090 B2
## bucket2008.fctr bucket2009.fctr.predict.
## 307446 B1 B1
## 307449 B1 B1
## 307450 B1 B1
## 307451 B1 B1
## 307452 B1 B1
## 366459 B5 B2
## bucket2009.fctr.predict..accurate .label
## 307446 FALSE .307446
## 307449 FALSE .307449
## 307450 FALSE .307450
## 307451 FALSE .307451
## 307452 FALSE .307452
## 366459 TRUE .366459
replay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"data.training.all.prediction","model.final")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
## 2.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction data.new.prediction firing: model.selected
## 3.0000 3 0 2 1 0
## 3.0000 multiple enabled transitions: model.final data.training.all.prediction data.new.prediction firing: data.training.all.prediction
## 4.0000 5 0 1 1 1
## 4.0000 multiple enabled transitions: model.final data.training.all.prediction data.new.prediction firing: model.final
## 5.0000 4 0 0 2 1
glb_script_df <- rbind(glb_script_df,
data.frame(chunk_label="predict.data.new",
chunk_step_major=max(glb_script_df$chunk_step_major)+1,
chunk_step_minor=0,
elapsed=(proc.time() - glb_script_tm)["elapsed"]))
print(tail(glb_script_df, 2))## chunk_label chunk_step_major chunk_step_minor
## elapsed10 fit.data.training.all 6 1
## elapsed11 predict.data.new 7 0
## elapsed
## elapsed10 9789.672
## elapsed11 10395.511Step 7: predict data.new
if (glb_is_regression)
glb_newent_df[, glb_rsp_var_out] <- predict(glb_fin_mdl,
newdata=glb_newent_df, type="response")
if (glb_is_classification & glb_is_binomial) {
# Compute selected model predictions
if (any(class(glb_fin_mdl) %in% c("train"))) {
glb_newent_df[, paste0(glb_rsp_var_out, ".proba")] <-
predict(glb_fin_mdl, newdata=glb_newent_df, type="prob")[, 2]
} else if (any(class(glb_fin_mdl) %in% c("rpart", "randomForest"))) {
glb_newent_df[, paste0(glb_rsp_var_out, ".proba")] <-
predict(glb_fin_mdl, newdata=glb_newent_df, type="prob")[, 2]
} else if (class(glb_fin_mdl) == "glm") {
stop("not implemented yet")
glb_newent_df[, paste0(glb_rsp_var_out, ".proba")] <-
predict(glb_fin_mdl, newdata=glb_newent_df, type="response")
} else stop("not implemented yet")
if ((class(glb_newent_df[, glb_rsp_var]) != "factor") |
(length(levels(glb_newent_df[, glb_rsp_var])) != 2))
stop("expecting a factor with two levels:", glb_rsp_var)
glb_newent_df[, glb_rsp_var_out] <-
factor(levels(glb_newent_df[, glb_rsp_var])[
(glb_newent_df[, paste0(glb_rsp_var_out, ".proba")] >=
glb_clf_proba_threshold) * 1 + 1])
# Compute dummy model predictions
glb_newent_df[, paste0(glb_rsp_var, ".predictdmy.proba")] <-
predict(glb_dmy_mdl, newdata=glb_newent_df, type="prob")[, 2]
if ((class(glb_newent_df[, glb_rsp_var]) != "factor") |
(length(levels(glb_newent_df[, glb_rsp_var])) != 2))
stop("expecting a factor with two levels:", glb_rsp_var)
glb_newent_df[, paste0(glb_rsp_var, ".predictdmy")] <-
factor(levels(glb_newent_df[, glb_rsp_var])[
(glb_newent_df[, paste0(glb_rsp_var, ".predictdmy.proba")] >=
glb_clf_proba_threshold) * 1 + 1])
}
if (glb_is_classification & !glb_is_binomial) {
# Compute final model predictions
glb_rsp_var_out <- paste0(glb_rsp_var_out, "Final")
glb_newent_df[, glb_rsp_var_out] <-
mypredict_mdl(glb_fin_mdl, glb_newent_df, glb_rsp_var, glb_rsp_var_out,
"Final", "Final",
glb_model_metric_smmry, glb_model_metric,
glb_model_metric_maximize, ret_type="raw")
}## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114141 8610 124 103 0
## B2 18409 16102 187 142 0
## B3 8027 8146 118 99 0
## B4 3099 4584 53 201 0
## B5 351 657 4 45 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.126669e-01 NA 7.105887e-01 7.147384e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 7.015238e-319 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflowmyprint_df(glb_newent_df[, c(glb_id_vars, glb_rsp_var, glb_rsp_var_out)])## bucket2009.fctr bucket2009.fctr.predict.Final
## 3 B1 B1
## 5 B1 B1
## 6 B1 B1
## 8 B1 B1
## 9 B1 B1
## 10 B1 B1
## bucket2009.fctr bucket2009.fctr.predict.Final
## 41663 B1 B1
## 152431 B1 B1
## 189586 B1 B1
## 254336 B1 B2
## 310981 B2 B1
## 350192 B2 B1
## bucket2009.fctr bucket2009.fctr.predict.Final
## 457996 B5 B2
## 457998 B5 B2
## 458001 B5 B2
## 458002 B5 B2
## 458003 B5 B2
## 458005 B5 B2if (glb_is_regression) {
print(sprintf("Total SSE: %0.4f",
sum((glb_newent_df[, glb_rsp_var_out] -
glb_newent_df[, glb_rsp_var]) ^ 2)))
print(sprintf("RMSE: %0.4f",
(sum((glb_newent_df[, glb_rsp_var_out] -
glb_newent_df[, glb_rsp_var]) ^ 2) / nrow(glb_newent_df)) ^ 0.5))
print(myplot_scatter(glb_newent_df, glb_rsp_var, glb_rsp_var_out,
smooth=TRUE))
glb_newent_df[, paste0(glb_rsp_var_out, ".err")] <-
abs(glb_newent_df[, glb_rsp_var_out] - glb_newent_df[, glb_rsp_var])
print(head(orderBy(reformulate(c("-", paste0(glb_rsp_var_out, ".err"))),
glb_newent_df)))
# glb_newent_df[, "<Output Pred variable>"] <- func(glb_newent_df[, glb_pred_var_name])
}
if (glb_is_classification & glb_is_binomial) {
ROCRpred <- prediction(glb_newent_df[, paste0(glb_rsp_var_out, ".proba")],
glb_newent_df[, glb_rsp_var])
print(sprintf("auc=%0.4f", auc <- as.numeric(performance(ROCRpred, "auc")@y.values)))
print(sprintf("probability threshold=%0.4f", glb_clf_proba_threshold))
print(newent_conf_df <- mycreate_xtab(glb_newent_df,
c(glb_rsp_var, glb_rsp_var_out)))
print(sprintf("f.score.sel=%0.4f",
mycompute_classifier_f.score(mdl=glb_fin_mdl, obs_df=glb_newent_df,
proba_threshold=glb_clf_proba_threshold,
rsp_var=glb_rsp_var,
rsp_var_out=glb_rsp_var_out)))
print(sprintf("sensitivity=%0.4f", newent_conf_df[2, 3] /
(newent_conf_df[2, 3] + newent_conf_df[2, 2])))
print(sprintf("specificity=%0.4f", newent_conf_df[1, 2] /
(newent_conf_df[1, 2] + newent_conf_df[1, 3])))
print(sprintf("accuracy=%0.4f", (newent_conf_df[1, 2] + newent_conf_df[2, 3]) /
(newent_conf_df[1, 2] + newent_conf_df[2, 3] +
newent_conf_df[1, 3] + newent_conf_df[2, 2])))
print(mycreate_xtab(glb_newent_df, c(glb_rsp_var, paste0(glb_rsp_var, ".predictdmy"))))
print(sprintf("f.score.dmy=%0.4f",
mycompute_classifier_f.score(mdl=glb_dmy_mdl, obs_df=glb_newent_df,
proba_threshold=glb_clf_proba_threshold,
rsp_var=glb_rsp_var,
rsp_var_out=paste0(glb_rsp_var, ".predictdmy"))))
}
if (glb_is_classification & !glb_is_binomial) {
print(mypredict_mdl(glb_fin_mdl, glb_newent_df, glb_rsp_var, glb_rsp_var_out,
"Final", "Final",
glb_model_metric_smmry, glb_model_metric,
glb_model_metric_maximize, ret_type="stats"))
} ## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## Prediction
## Reference B1 B2 B3 B4 B5
## B1 114141 8610 124 103 0
## B2 18409 16102 187 142 0
## B3 8027 8146 118 99 0
## B4 3099 4584 53 201 0
## B5 351 657 4 45 0
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.126669e-01 NA 7.105887e-01 7.147384e-01 6.712700e-01
## AccuracyPValue McnemarPValue
## 7.015238e-319 0.000000e+00## Warning in ni[1:m] * nj[1:m]: NAs produced by integer overflow## model_id max.Accuracy.Final max.AccuracyLower.Final
## 1 Final 0.7126669 0.7105887
## max.AccuracyUpper.Final max.Kappa.Final min.loss.error.Final
## 1 0.7147384 NA 0.7578902glb_analytics_diag_plots(obs_df=glb_newent_df)
## age alzheimers arthritis cancer copd depression diabetes
## 3 67 0 0 0 0 0 0
## 89984 72 0 1 0 0 0 1
## 89990 86 0 0 0 0 1 1
## 90000 80 0 0 0 0 0 1
## 90002 76 0 1 0 0 0 1
## 90008 79 0 0 0 0 1 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 3 0 0 0 0 0 0
## 89984 1 1 1 0 0 2420
## 89990 1 1 1 0 0 2530
## 90000 1 1 1 0 0 2720
## 90002 1 1 1 0 0 2730
## 90008 1 1 0 0 1 2900
## bucket2008 reimbursement2009 bucket2009 .rnorm bucket2009.fctr
## 3 1 0 1 0.7183313 B1
## 89984 1 0 1 1.9883004 B1
## 89990 1 0 1 -0.5157387 B1
## 90000 1 0 1 0.8442956 B1
## 90002 1 0 1 -0.8136275 B1
## 90008 1 0 1 -0.3684009 B1
## bucket2008.fctr bucket2009.fctr.predict.Final
## 3 B1 B1
## 89984 B1 B2
## 89990 B1 B2
## 90000 B1 B2
## 90002 B1 B2
## 90008 B1 B2
## bucket2009.fctr.predict.Final.accurate .label
## 3 TRUE .3
## 89984 FALSE .89984
## 89990 FALSE .89990
## 90000 FALSE .90000
## 90002 FALSE .90002
## 90008 FALSE .90008
## age alzheimers arthritis cancer copd depression diabetes
## 89984 72 0 1 0 0 0 1
## 89990 86 0 0 0 0 1 1
## 90000 80 0 0 0 0 0 1
## 125189 84 0 0 1 0 1 1
## 307454 66 0 0 0 0 0 0
## 307456 99 0 0 0 0 0 0
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 89984 1 1 1 0 0 2420
## 89990 1 1 1 0 0 2530
## 90000 1 1 1 0 0 2720
## 125189 1 1 1 0 0 62830
## 307454 0 0 0 0 0 0
## 307456 0 0 0 0 0 0
## bucket2008 reimbursement2009 bucket2009 .rnorm bucket2009.fctr
## 89984 1 0 1 1.9883004 B1
## 89990 1 0 1 -0.5157387 B1
## 90000 1 0 1 0.8442956 B1
## 125189 5 210 1 0.8591767 B1
## 307454 1 3000 2 -0.3857694 B2
## 307456 1 3000 2 0.6728828 B2
## bucket2008.fctr bucket2009.fctr.predict.Final
## 89984 B1 B2
## 89990 B1 B2
## 90000 B1 B2
## 125189 B5 B2
## 307454 B1 B1
## 307456 B1 B1
## bucket2009.fctr.predict.Final.accurate .label
## 89984 FALSE .89984
## 89990 FALSE .89990
## 90000 FALSE .90000
## 125189 FALSE .125189
## 307454 FALSE .307454
## 307456 FALSE .307456
## age alzheimers arthritis cancer copd depression diabetes
## 307448 80 0 0 0 0 0 0
## 307454 66 0 0 0 0 0 0
## 307455 66 0 0 0 0 0 0
## 307456 99 0 0 0 0 0 0
## 307457 66 0 0 0 0 0 0
## 449953 80 1 0 0 1 1 1
## heart.failure ihd kidney osteoporosis stroke reimbursement2008
## 307448 0 0 0 0 0 0
## 307454 0 0 0 0 0 0
## 307455 0 0 0 0 0 0
## 307456 0 0 0 0 0 0
## 307457 0 0 0 0 0 0
## 449953 1 1 1 0 1 221640
## bucket2008 reimbursement2009 bucket2009 .rnorm bucket2009.fctr
## 307448 1 3000 2 -1.2520078 B2
## 307454 1 3000 2 -0.3857694 B2
## 307455 1 3000 2 0.2479826 B2
## 307456 1 3000 2 0.6728828 B2
## 307457 1 3000 2 0.3236726 B2
## 449953 5 34130 4 -0.7428332 B4
## bucket2008.fctr bucket2009.fctr.predict.Final
## 307448 B1 B1
## 307454 B1 B1
## 307455 B1 B1
## 307456 B1 B1
## 307457 B1 B1
## 449953 B5 B2
## bucket2009.fctr.predict.Final.accurate .label
## 307448 FALSE .307448
## 307454 FALSE .307454
## 307455 FALSE .307455
## 307456 FALSE .307456
## 307457 FALSE .307457
## 449953 FALSE .449953
tmp_replay_lst <- replay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"data.new.prediction")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
## 2.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction data.new.prediction firing: model.selected
## 3.0000 3 0 2 1 0
## 3.0000 multiple enabled transitions: model.final data.training.all.prediction data.new.prediction firing: data.training.all.prediction
## 4.0000 5 0 1 1 1
## 4.0000 multiple enabled transitions: model.final data.training.all.prediction data.new.prediction firing: model.final
## 5.0000 4 0 0 2 1
## 6.0000 6 0 0 1 2
#print(ggplot.petrinet(tmp_replay_lst[["pn"]]) + coord_flip())Null Hypothesis (\(\sf{H_{0}}\)): mpg is not impacted by am_fctr.
The variance by am_fctr appears to be independent. #{r q1, cache=FALSE} # print(t.test(subset(cars_df, am_fctr == "automatic")$mpg, # subset(cars_df, am_fctr == "manual")$mpg, # var.equal=FALSE)$conf) # We reject the null hypothesis i.e. we have evidence to conclude that am_fctr impacts mpg (95% confidence). Manual transmission is better for miles per gallon versus automatic transmission.
## chunk_label chunk_step_major chunk_step_minor elapsed
## 10 fit.data.training.all 6 0 9768.275
## 12 predict.data.new 7 0 10395.511
## 9 fit.models 5 0 71.848
## 11 fit.data.training.all 6 1 9789.672
## 6 extract_features 3 0 26.774
## 4 manage_missing_data 2 2 16.418
## 2 cleanse_data 2 0 7.036
## 7 select_features 4 0 28.472
## 8 remove_correlated_features 4 1 29.537
## 5 encode_retype_data 2 3 16.992
## 3 inspectORexplore.data 2 1 7.066
## 1 import_data 1 0 0.003
## elapsed_diff
## 10 9696.427
## 12 605.839
## 9 42.311
## 11 21.397
## 6 9.782
## 4 9.352
## 2 7.033
## 7 1.698
## 8 1.065
## 5 0.574
## 3 0.030
## 1 0.000## [1] "Total Elapsed Time: 10,395.51 secs"
## R version 3.1.3 (2015-03-09)
## Platform: x86_64-apple-darwin13.4.0 (64-bit)
## Running under: OS X 10.10.3 (Yosemite)
##
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## attached base packages:
## [1] tcltk grid stats graphics grDevices utils datasets
## [8] methods base
##
## other attached packages:
## [1] randomForest_4.6-10 rpart.plot_1.5.2 rpart_4.1-9
## [4] caret_6.0-41 lattice_0.20-31 reshape2_1.4.1
## [7] sqldf_0.4-10 RSQLite_1.0.0 DBI_0.3.1
## [10] gsubfn_0.6-6 proto_0.3-10 plyr_1.8.1
## [13] caTools_1.17.1 doBy_4.5-13 survival_2.38-1
## [16] ggplot2_1.0.1
##
## loaded via a namespace (and not attached):
## [1] bitops_1.0-6 BradleyTerry2_1.0-6 brglm_0.5-9
## [4] car_2.0-25 chron_2.3-45 class_7.3-12
## [7] codetools_0.2-11 colorspace_1.2-6 compiler_3.1.3
## [10] digest_0.6.8 e1071_1.6-4 evaluate_0.5.5
## [13] foreach_1.4.2 formatR_1.1 gtable_0.1.2
## [16] gtools_3.4.1 htmltools_0.2.6 iterators_1.0.7
## [19] knitr_1.9 labeling_0.3 lme4_1.1-7
## [22] MASS_7.3-40 Matrix_1.2-0 mgcv_1.8-6
## [25] minqa_1.2.4 munsell_0.4.2 nlme_3.1-120
## [28] nloptr_1.0.4 nnet_7.3-9 parallel_3.1.3
## [31] pbkrtest_0.4-2 quantreg_5.11 RColorBrewer_1.1-2
## [34] Rcpp_0.11.5 rmarkdown_0.5.1 scales_0.2.4
## [37] SparseM_1.6 splines_3.1.3 stringr_0.6.2
## [40] tools_3.1.3 yaml_2.1.13